[clang] 3de645e - [OpenMP][NFC] Split the reduction buffer size into two components
Johannes Doerfert via cfe-commits
cfe-commits at lists.llvm.org
Mon Nov 6 11:51:50 PST 2023
Author: Johannes Doerfert
Date: 2023-11-06T11:50:41-08:00
New Revision: 3de645efe30b83ba1b6d7e500486c4f441a17a61
URL: https://github.com/llvm/llvm-project/commit/3de645efe30b83ba1b6d7e500486c4f441a17a61
DIFF: https://github.com/llvm/llvm-project/commit/3de645efe30b83ba1b6d7e500486c4f441a17a61.diff
LOG: [OpenMP][NFC] Split the reduction buffer size into two components
Before we tracked the size of the teams reduction buffer in order to
allocate it at runtime per kernel launch. This patch splits the number
into two parts, the size of the reduction data (=all reduction
variables) and the (maximal) length of the buffer. This will allow us to
allocate less if we need less, e.g., if we have less teams than the
maximal length. It also allows us to move code from clangs codegen into
the runtime as we now know how large the reduction data is.
Added:
Modified:
clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp
clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp
clang/test/OpenMP/reduction_implicit_map.cpp
clang/test/OpenMP/target_teams_generic_loop_codegen.cpp
llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
llvm/test/Transforms/OpenMP/add_attributes.ll
llvm/test/Transforms/OpenMP/always_inline_device.ll
llvm/test/Transforms/OpenMP/custom_state_machines.ll
llvm/test/Transforms/OpenMP/custom_state_machines_pre_lto.ll
llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll
llvm/test/Transforms/OpenMP/deduplication_target.ll
llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll
llvm/test/Transforms/OpenMP/global_constructor.ll
llvm/test/Transforms/OpenMP/globalization_remarks.ll
llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll
llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll
llvm/test/Transforms/OpenMP/nested_parallelism.ll
llvm/test/Transforms/OpenMP/parallel_level_fold.ll
llvm/test/Transforms/OpenMP/remove_globalization.ll
llvm/test/Transforms/OpenMP/replace_globalization.ll
llvm/test/Transforms/OpenMP/single_threaded_execution.ll
llvm/test/Transforms/OpenMP/spmdization.ll
llvm/test/Transforms/OpenMP/spmdization_assumes.ll
llvm/test/Transforms/OpenMP/spmdization_constant_prop.ll
llvm/test/Transforms/OpenMP/spmdization_guarding.ll
llvm/test/Transforms/OpenMP/spmdization_guarding_two_reaching_kernels.ll
llvm/test/Transforms/OpenMP/spmdization_indirect.ll
llvm/test/Transforms/OpenMP/spmdization_kernel_env_dep.ll
llvm/test/Transforms/OpenMP/spmdization_no_guarding_two_reaching_kernels.ll
llvm/test/Transforms/OpenMP/spmdization_remarks.ll
llvm/test/Transforms/OpenMP/value-simplify-openmp-opt.ll
llvm/test/Transforms/PhaseOrdering/openmp-opt-module.ll
mlir/test/Target/LLVMIR/omptarget-region-device-llvm.mlir
openmp/libomptarget/DeviceRTL/include/Interface.h
openmp/libomptarget/DeviceRTL/src/Reduction.cpp
openmp/libomptarget/include/Environment.h
openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp
Removed:
################################################################################
diff --git a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
index b4e067ff497a085..a13d74743c3bd3f 100644
--- a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
+++ b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
@@ -801,10 +801,12 @@ void CGOpenMPRuntimeGPU::emitKernelDeinit(CodeGenFunction &CGF,
llvm::Type *LLVMReductionsBufferTy =
CGM.getTypes().ConvertTypeForMem(StaticTy);
const auto &DL = CGM.getModule().getDataLayout();
- uint64_t BufferSize =
+ uint64_t ReductionDataSize =
DL.getTypeAllocSize(LLVMReductionsBufferTy).getFixedValue();
CGBuilderTy &Bld = CGF.Builder;
- OMPBuilder.createTargetDeinit(Bld, BufferSize);
+ OMPBuilder.createTargetDeinit(Bld, ReductionDataSize,
+ C.getLangOpts().OpenMPCUDAReductionBufNum);
+ TeamsReductions.clear();
}
void CGOpenMPRuntimeGPU::emitSPMDKernel(const OMPExecutableDirective &D,
@@ -2828,15 +2830,25 @@ void CGOpenMPRuntimeGPU::emitReduction(
assert((TeamsReduction || ParallelReduction) &&
"Invalid reduction selection in emitReduction.");
+ llvm::SmallDenseMap<const ValueDecl *, const FieldDecl *> VarFieldMap;
+ llvm::SmallVector<const ValueDecl *, 4> PrivatesReductions(Privates.size());
+ int Cnt = 0;
+ for (const Expr *DRE : Privates) {
+ PrivatesReductions[Cnt] = cast<DeclRefExpr>(DRE)->getDecl();
+ ++Cnt;
+ }
+
+ ASTContext &C = CGM.getContext();
+ const RecordDecl *ReductionRec = ::buildRecordForGlobalizedVars(
+ CGM.getContext(), PrivatesReductions, std::nullopt, VarFieldMap, 1);
+
// Build res = __kmpc_reduce{_nowait}(<gtid>, <n>, sizeof(RedList),
// RedList, shuffle_reduce_func, interwarp_copy_func);
// or
// Build res = __kmpc_reduce_teams_nowait_simple(<loc>, <gtid>, <lck>);
llvm::Value *RTLoc = emitUpdateLocation(CGF, Loc);
- llvm::Value *ThreadId = getThreadID(CGF, Loc);
llvm::Value *Res;
- ASTContext &C = CGM.getContext();
// 1. Build a list of reduction variables.
// void *RedList[<n>] = {<ReductionVars>[0], ..., <ReductionVars>[<n>-1]};
auto Size = RHSExprs.size();
@@ -2878,19 +2890,17 @@ void CGOpenMPRuntimeGPU::emitReduction(
llvm::Function *ReductionFn = emitReductionFunction(
CGF.CurFn->getName(), Loc, CGF.ConvertTypeForMem(ReductionArrayTy),
Privates, LHSExprs, RHSExprs, ReductionOps);
- llvm::Value *ReductionArrayTySize = CGF.getTypeSize(ReductionArrayTy);
+ llvm::Value *ReductionDataSize =
+ CGF.getTypeSize(C.getRecordType(ReductionRec));
+ ReductionDataSize =
+ CGF.Builder.CreateSExtOrTrunc(ReductionDataSize, CGF.Int64Ty);
llvm::Function *ShuffleAndReduceFn = emitShuffleAndReduceFunction(
CGM, Privates, ReductionArrayTy, ReductionFn, Loc);
llvm::Value *InterWarpCopyFn =
emitInterWarpCopyFunction(CGM, Privates, ReductionArrayTy, Loc);
if (ParallelReduction) {
- llvm::Value *Args[] = {RTLoc,
- ThreadId,
- CGF.Builder.getInt32(RHSExprs.size()),
- ReductionArrayTySize,
- RL,
- ShuffleAndReduceFn,
+ llvm::Value *Args[] = {RTLoc, ReductionDataSize, RL, ShuffleAndReduceFn,
InterWarpCopyFn};
Res = CGF.EmitRuntimeCall(
@@ -2899,37 +2909,27 @@ void CGOpenMPRuntimeGPU::emitReduction(
Args);
} else {
assert(TeamsReduction && "expected teams reduction.");
- llvm::SmallDenseMap<const ValueDecl *, const FieldDecl *> VarFieldMap;
- llvm::SmallVector<const ValueDecl *, 4> PrivatesReductions(Privates.size());
- int Cnt = 0;
- for (const Expr *DRE : Privates) {
- PrivatesReductions[Cnt] = cast<DeclRefExpr>(DRE)->getDecl();
- ++Cnt;
- }
- const RecordDecl *TeamReductionRec = ::buildRecordForGlobalizedVars(
- CGM.getContext(), PrivatesReductions, std::nullopt, VarFieldMap,
- C.getLangOpts().OpenMPCUDAReductionBufNum);
- TeamsReductions.push_back(TeamReductionRec);
+ TeamsReductions.push_back(ReductionRec);
auto *KernelTeamsReductionPtr = CGF.EmitRuntimeCall(
OMPBuilder.getOrCreateRuntimeFunction(
CGM.getModule(), OMPRTL___kmpc_reduction_get_fixed_buffer),
{}, "_openmp_teams_reductions_buffer_$_$ptr");
llvm::Value *GlobalToBufferCpyFn = ::emitListToGlobalCopyFunction(
- CGM, Privates, ReductionArrayTy, Loc, TeamReductionRec, VarFieldMap);
+ CGM, Privates, ReductionArrayTy, Loc, ReductionRec, VarFieldMap);
llvm::Value *GlobalToBufferRedFn = ::emitListToGlobalReduceFunction(
- CGM, Privates, ReductionArrayTy, Loc, TeamReductionRec, VarFieldMap,
+ CGM, Privates, ReductionArrayTy, Loc, ReductionRec, VarFieldMap,
ReductionFn);
llvm::Value *BufferToGlobalCpyFn = ::emitGlobalToListCopyFunction(
- CGM, Privates, ReductionArrayTy, Loc, TeamReductionRec, VarFieldMap);
+ CGM, Privates, ReductionArrayTy, Loc, ReductionRec, VarFieldMap);
llvm::Value *BufferToGlobalRedFn = ::emitGlobalToListReduceFunction(
- CGM, Privates, ReductionArrayTy, Loc, TeamReductionRec, VarFieldMap,
+ CGM, Privates, ReductionArrayTy, Loc, ReductionRec, VarFieldMap,
ReductionFn);
llvm::Value *Args[] = {
RTLoc,
- ThreadId,
KernelTeamsReductionPtr,
CGF.Builder.getInt32(C.getLangOpts().OpenMPCUDAReductionBufNum),
+ ReductionDataSize,
RL,
ShuffleAndReduceFn,
InterWarpCopyFn,
diff --git a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
index c2a958dfdd2453e..c077b58f614fa14 100644
--- a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
@@ -136,17 +136,15 @@ int bar(int n){
// CHECK-64-NEXT: [[TMP1:%.*]] = load double, ptr [[E1]], align 8
// CHECK-64-NEXT: [[ADD:%.*]] = fadd double [[TMP1]], 5.000000e+00
// CHECK-64-NEXT: store double [[ADD]], ptr [[E1]], align 8
-// CHECK-64-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK-64-NEXT: [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
-// CHECK-64-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK-64-NEXT: store ptr [[E1]], ptr [[TMP4]], align 8
-// CHECK-64-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP3]], i32 1, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
-// CHECK-64-NEXT: [[TMP6:%.*]] = icmp eq i32 [[TMP5]], 1
-// CHECK-64-NEXT: br i1 [[TMP6]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-64-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK-64-NEXT: store ptr [[E1]], ptr [[TMP2]], align 8
+// CHECK-64-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
+// CHECK-64-NEXT: [[TMP4:%.*]] = icmp eq i32 [[TMP3]], 1
+// CHECK-64-NEXT: br i1 [[TMP4]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK-64: .omp.reduction.then:
-// CHECK-64-NEXT: [[TMP7:%.*]] = load double, ptr [[TMP0]], align 8
-// CHECK-64-NEXT: [[TMP8:%.*]] = load double, ptr [[E1]], align 8
-// CHECK-64-NEXT: [[ADD2:%.*]] = fadd double [[TMP7]], [[TMP8]]
+// CHECK-64-NEXT: [[TMP5:%.*]] = load double, ptr [[TMP0]], align 8
+// CHECK-64-NEXT: [[TMP6:%.*]] = load double, ptr [[E1]], align 8
+// CHECK-64-NEXT: [[ADD2:%.*]] = fadd double [[TMP5]], [[TMP6]]
// CHECK-64-NEXT: store double [[ADD2]], ptr [[TMP0]], align 8
// CHECK-64-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK-64: .omp.reduction.done:
@@ -331,26 +329,24 @@ int bar(int n){
// CHECK-64-NEXT: [[TMP3:%.*]] = load float, ptr [[D2]], align 4
// CHECK-64-NEXT: [[MUL:%.*]] = fmul float [[TMP3]], 3.300000e+01
// CHECK-64-NEXT: store float [[MUL]], ptr [[D2]], align 4
-// CHECK-64-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK-64-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP4]], align 4
-// CHECK-64-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK-64-NEXT: store ptr [[C1]], ptr [[TMP6]], align 8
-// CHECK-64-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK-64-NEXT: store ptr [[D2]], ptr [[TMP7]], align 8
-// CHECK-64-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP5]], i32 2, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2)
-// CHECK-64-NEXT: [[TMP9:%.*]] = icmp eq i32 [[TMP8]], 1
-// CHECK-64-NEXT: br i1 [[TMP9]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-64-NEXT: [[TMP4:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK-64-NEXT: store ptr [[C1]], ptr [[TMP4]], align 8
+// CHECK-64-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK-64-NEXT: store ptr [[D2]], ptr [[TMP5]], align 8
+// CHECK-64-NEXT: [[TMP6:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2)
+// CHECK-64-NEXT: [[TMP7:%.*]] = icmp eq i32 [[TMP6]], 1
+// CHECK-64-NEXT: br i1 [[TMP7]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK-64: .omp.reduction.then:
-// CHECK-64-NEXT: [[TMP10:%.*]] = load i8, ptr [[TMP0]], align 1
-// CHECK-64-NEXT: [[CONV4:%.*]] = sext i8 [[TMP10]] to i32
-// CHECK-64-NEXT: [[TMP11:%.*]] = load i8, ptr [[C1]], align 1
-// CHECK-64-NEXT: [[CONV5:%.*]] = sext i8 [[TMP11]] to i32
+// CHECK-64-NEXT: [[TMP8:%.*]] = load i8, ptr [[TMP0]], align 1
+// CHECK-64-NEXT: [[CONV4:%.*]] = sext i8 [[TMP8]] to i32
+// CHECK-64-NEXT: [[TMP9:%.*]] = load i8, ptr [[C1]], align 1
+// CHECK-64-NEXT: [[CONV5:%.*]] = sext i8 [[TMP9]] to i32
// CHECK-64-NEXT: [[XOR6:%.*]] = xor i32 [[CONV4]], [[CONV5]]
// CHECK-64-NEXT: [[CONV7:%.*]] = trunc i32 [[XOR6]] to i8
// CHECK-64-NEXT: store i8 [[CONV7]], ptr [[TMP0]], align 1
-// CHECK-64-NEXT: [[TMP12:%.*]] = load float, ptr [[TMP1]], align 4
-// CHECK-64-NEXT: [[TMP13:%.*]] = load float, ptr [[D2]], align 4
-// CHECK-64-NEXT: [[MUL8:%.*]] = fmul float [[TMP12]], [[TMP13]]
+// CHECK-64-NEXT: [[TMP10:%.*]] = load float, ptr [[TMP1]], align 4
+// CHECK-64-NEXT: [[TMP11:%.*]] = load float, ptr [[D2]], align 4
+// CHECK-64-NEXT: [[MUL8:%.*]] = fmul float [[TMP10]], [[TMP11]]
// CHECK-64-NEXT: store float [[MUL8]], ptr [[TMP1]], align 4
// CHECK-64-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK-64: .omp.reduction.done:
@@ -578,34 +574,32 @@ int bar(int n){
// CHECK-64-NEXT: [[COND:%.*]] = phi i32 [ 99, [[COND_TRUE]] ], [ [[CONV3]], [[COND_FALSE]] ]
// CHECK-64-NEXT: [[CONV4:%.*]] = trunc i32 [[COND]] to i16
// CHECK-64-NEXT: store i16 [[CONV4]], ptr [[B2]], align 2
-// CHECK-64-NEXT: [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK-64-NEXT: [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
-// CHECK-64-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK-64-NEXT: store ptr [[A1]], ptr [[TMP7]], align 8
-// CHECK-64-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK-64-NEXT: store ptr [[B2]], ptr [[TMP8]], align 8
-// CHECK-64-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP6]], i32 2, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func3, ptr @_omp_reduction_inter_warp_copy_func4)
-// CHECK-64-NEXT: [[TMP10:%.*]] = icmp eq i32 [[TMP9]], 1
-// CHECK-64-NEXT: br i1 [[TMP10]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-64-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK-64-NEXT: store ptr [[A1]], ptr [[TMP5]], align 8
+// CHECK-64-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK-64-NEXT: store ptr [[B2]], ptr [[TMP6]], align 8
+// CHECK-64-NEXT: [[TMP7:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func3, ptr @_omp_reduction_inter_warp_copy_func4)
+// CHECK-64-NEXT: [[TMP8:%.*]] = icmp eq i32 [[TMP7]], 1
+// CHECK-64-NEXT: br i1 [[TMP8]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK-64: .omp.reduction.then:
-// CHECK-64-NEXT: [[TMP11:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK-64-NEXT: [[TMP12:%.*]] = load i32, ptr [[A1]], align 4
-// CHECK-64-NEXT: [[OR5:%.*]] = or i32 [[TMP11]], [[TMP12]]
+// CHECK-64-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK-64-NEXT: [[TMP10:%.*]] = load i32, ptr [[A1]], align 4
+// CHECK-64-NEXT: [[OR5:%.*]] = or i32 [[TMP9]], [[TMP10]]
// CHECK-64-NEXT: store i32 [[OR5]], ptr [[TMP0]], align 4
-// CHECK-64-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
-// CHECK-64-NEXT: [[CONV6:%.*]] = sext i16 [[TMP13]] to i32
-// CHECK-64-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
-// CHECK-64-NEXT: [[CONV7:%.*]] = sext i16 [[TMP14]] to i32
+// CHECK-64-NEXT: [[TMP11:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK-64-NEXT: [[CONV6:%.*]] = sext i16 [[TMP11]] to i32
+// CHECK-64-NEXT: [[TMP12:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK-64-NEXT: [[CONV7:%.*]] = sext i16 [[TMP12]] to i32
// CHECK-64-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[CONV6]], [[CONV7]]
// CHECK-64-NEXT: br i1 [[CMP8]], label [[COND_TRUE9:%.*]], label [[COND_FALSE10:%.*]]
// CHECK-64: cond.true9:
-// CHECK-64-NEXT: [[TMP15:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK-64-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
// CHECK-64-NEXT: br label [[COND_END11:%.*]]
// CHECK-64: cond.false10:
-// CHECK-64-NEXT: [[TMP16:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK-64-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
// CHECK-64-NEXT: br label [[COND_END11]]
// CHECK-64: cond.end11:
-// CHECK-64-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP15]], [[COND_TRUE9]] ], [ [[TMP16]], [[COND_FALSE10]] ]
+// CHECK-64-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP13]], [[COND_TRUE9]] ], [ [[TMP14]], [[COND_FALSE10]] ]
// CHECK-64-NEXT: store i16 [[COND12]], ptr [[TMP1]], align 2
// CHECK-64-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK-64: .omp.reduction.done:
@@ -809,17 +803,15 @@ int bar(int n){
// CHECK-32-NEXT: [[TMP1:%.*]] = load double, ptr [[E1]], align 8
// CHECK-32-NEXT: [[ADD:%.*]] = fadd double [[TMP1]], 5.000000e+00
// CHECK-32-NEXT: store double [[ADD]], ptr [[E1]], align 8
-// CHECK-32-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK-32-NEXT: [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
-// CHECK-32-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK-32-NEXT: store ptr [[E1]], ptr [[TMP4]], align 4
-// CHECK-32-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP3]], i32 1, i32 4, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
-// CHECK-32-NEXT: [[TMP6:%.*]] = icmp eq i32 [[TMP5]], 1
-// CHECK-32-NEXT: br i1 [[TMP6]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-32-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK-32-NEXT: store ptr [[E1]], ptr [[TMP2]], align 4
+// CHECK-32-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
+// CHECK-32-NEXT: [[TMP4:%.*]] = icmp eq i32 [[TMP3]], 1
+// CHECK-32-NEXT: br i1 [[TMP4]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK-32: .omp.reduction.then:
-// CHECK-32-NEXT: [[TMP7:%.*]] = load double, ptr [[TMP0]], align 8
-// CHECK-32-NEXT: [[TMP8:%.*]] = load double, ptr [[E1]], align 8
-// CHECK-32-NEXT: [[ADD2:%.*]] = fadd double [[TMP7]], [[TMP8]]
+// CHECK-32-NEXT: [[TMP5:%.*]] = load double, ptr [[TMP0]], align 8
+// CHECK-32-NEXT: [[TMP6:%.*]] = load double, ptr [[E1]], align 8
+// CHECK-32-NEXT: [[ADD2:%.*]] = fadd double [[TMP5]], [[TMP6]]
// CHECK-32-NEXT: store double [[ADD2]], ptr [[TMP0]], align 8
// CHECK-32-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK-32: .omp.reduction.done:
@@ -1004,26 +996,24 @@ int bar(int n){
// CHECK-32-NEXT: [[TMP3:%.*]] = load float, ptr [[D2]], align 4
// CHECK-32-NEXT: [[MUL:%.*]] = fmul float [[TMP3]], 3.300000e+01
// CHECK-32-NEXT: store float [[MUL]], ptr [[D2]], align 4
-// CHECK-32-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK-32-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP4]], align 4
-// CHECK-32-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK-32-NEXT: store ptr [[C1]], ptr [[TMP6]], align 4
-// CHECK-32-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
-// CHECK-32-NEXT: store ptr [[D2]], ptr [[TMP7]], align 4
-// CHECK-32-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP5]], i32 2, i32 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2)
-// CHECK-32-NEXT: [[TMP9:%.*]] = icmp eq i32 [[TMP8]], 1
-// CHECK-32-NEXT: br i1 [[TMP9]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-32-NEXT: [[TMP4:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK-32-NEXT: store ptr [[C1]], ptr [[TMP4]], align 4
+// CHECK-32-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
+// CHECK-32-NEXT: store ptr [[D2]], ptr [[TMP5]], align 4
+// CHECK-32-NEXT: [[TMP6:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2)
+// CHECK-32-NEXT: [[TMP7:%.*]] = icmp eq i32 [[TMP6]], 1
+// CHECK-32-NEXT: br i1 [[TMP7]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK-32: .omp.reduction.then:
-// CHECK-32-NEXT: [[TMP10:%.*]] = load i8, ptr [[TMP0]], align 1
-// CHECK-32-NEXT: [[CONV4:%.*]] = sext i8 [[TMP10]] to i32
-// CHECK-32-NEXT: [[TMP11:%.*]] = load i8, ptr [[C1]], align 1
-// CHECK-32-NEXT: [[CONV5:%.*]] = sext i8 [[TMP11]] to i32
+// CHECK-32-NEXT: [[TMP8:%.*]] = load i8, ptr [[TMP0]], align 1
+// CHECK-32-NEXT: [[CONV4:%.*]] = sext i8 [[TMP8]] to i32
+// CHECK-32-NEXT: [[TMP9:%.*]] = load i8, ptr [[C1]], align 1
+// CHECK-32-NEXT: [[CONV5:%.*]] = sext i8 [[TMP9]] to i32
// CHECK-32-NEXT: [[XOR6:%.*]] = xor i32 [[CONV4]], [[CONV5]]
// CHECK-32-NEXT: [[CONV7:%.*]] = trunc i32 [[XOR6]] to i8
// CHECK-32-NEXT: store i8 [[CONV7]], ptr [[TMP0]], align 1
-// CHECK-32-NEXT: [[TMP12:%.*]] = load float, ptr [[TMP1]], align 4
-// CHECK-32-NEXT: [[TMP13:%.*]] = load float, ptr [[D2]], align 4
-// CHECK-32-NEXT: [[MUL8:%.*]] = fmul float [[TMP12]], [[TMP13]]
+// CHECK-32-NEXT: [[TMP10:%.*]] = load float, ptr [[TMP1]], align 4
+// CHECK-32-NEXT: [[TMP11:%.*]] = load float, ptr [[D2]], align 4
+// CHECK-32-NEXT: [[MUL8:%.*]] = fmul float [[TMP10]], [[TMP11]]
// CHECK-32-NEXT: store float [[MUL8]], ptr [[TMP1]], align 4
// CHECK-32-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK-32: .omp.reduction.done:
@@ -1251,34 +1241,32 @@ int bar(int n){
// CHECK-32-NEXT: [[COND:%.*]] = phi i32 [ 99, [[COND_TRUE]] ], [ [[CONV3]], [[COND_FALSE]] ]
// CHECK-32-NEXT: [[CONV4:%.*]] = trunc i32 [[COND]] to i16
// CHECK-32-NEXT: store i16 [[CONV4]], ptr [[B2]], align 2
-// CHECK-32-NEXT: [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK-32-NEXT: [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
-// CHECK-32-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK-32-NEXT: store ptr [[A1]], ptr [[TMP7]], align 4
-// CHECK-32-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
-// CHECK-32-NEXT: store ptr [[B2]], ptr [[TMP8]], align 4
-// CHECK-32-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP6]], i32 2, i32 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func3, ptr @_omp_reduction_inter_warp_copy_func4)
-// CHECK-32-NEXT: [[TMP10:%.*]] = icmp eq i32 [[TMP9]], 1
-// CHECK-32-NEXT: br i1 [[TMP10]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-32-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK-32-NEXT: store ptr [[A1]], ptr [[TMP5]], align 4
+// CHECK-32-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
+// CHECK-32-NEXT: store ptr [[B2]], ptr [[TMP6]], align 4
+// CHECK-32-NEXT: [[TMP7:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func3, ptr @_omp_reduction_inter_warp_copy_func4)
+// CHECK-32-NEXT: [[TMP8:%.*]] = icmp eq i32 [[TMP7]], 1
+// CHECK-32-NEXT: br i1 [[TMP8]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK-32: .omp.reduction.then:
-// CHECK-32-NEXT: [[TMP11:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK-32-NEXT: [[TMP12:%.*]] = load i32, ptr [[A1]], align 4
-// CHECK-32-NEXT: [[OR5:%.*]] = or i32 [[TMP11]], [[TMP12]]
+// CHECK-32-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK-32-NEXT: [[TMP10:%.*]] = load i32, ptr [[A1]], align 4
+// CHECK-32-NEXT: [[OR5:%.*]] = or i32 [[TMP9]], [[TMP10]]
// CHECK-32-NEXT: store i32 [[OR5]], ptr [[TMP0]], align 4
-// CHECK-32-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
-// CHECK-32-NEXT: [[CONV6:%.*]] = sext i16 [[TMP13]] to i32
-// CHECK-32-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
-// CHECK-32-NEXT: [[CONV7:%.*]] = sext i16 [[TMP14]] to i32
+// CHECK-32-NEXT: [[TMP11:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK-32-NEXT: [[CONV6:%.*]] = sext i16 [[TMP11]] to i32
+// CHECK-32-NEXT: [[TMP12:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK-32-NEXT: [[CONV7:%.*]] = sext i16 [[TMP12]] to i32
// CHECK-32-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[CONV6]], [[CONV7]]
// CHECK-32-NEXT: br i1 [[CMP8]], label [[COND_TRUE9:%.*]], label [[COND_FALSE10:%.*]]
// CHECK-32: cond.true9:
-// CHECK-32-NEXT: [[TMP15:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK-32-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
// CHECK-32-NEXT: br label [[COND_END11:%.*]]
// CHECK-32: cond.false10:
-// CHECK-32-NEXT: [[TMP16:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK-32-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
// CHECK-32-NEXT: br label [[COND_END11]]
// CHECK-32: cond.end11:
-// CHECK-32-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP15]], [[COND_TRUE9]] ], [ [[TMP16]], [[COND_FALSE10]] ]
+// CHECK-32-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP13]], [[COND_TRUE9]] ], [ [[TMP14]], [[COND_FALSE10]] ]
// CHECK-32-NEXT: store i16 [[COND12]], ptr [[TMP1]], align 2
// CHECK-32-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK-32: .omp.reduction.done:
@@ -1482,17 +1470,15 @@ int bar(int n){
// CHECK-32-EX-NEXT: [[TMP1:%.*]] = load double, ptr [[E1]], align 8
// CHECK-32-EX-NEXT: [[ADD:%.*]] = fadd double [[TMP1]], 5.000000e+00
// CHECK-32-EX-NEXT: store double [[ADD]], ptr [[E1]], align 8
-// CHECK-32-EX-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK-32-EX-NEXT: [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
-// CHECK-32-EX-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK-32-EX-NEXT: store ptr [[E1]], ptr [[TMP4]], align 4
-// CHECK-32-EX-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP3]], i32 1, i32 4, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
-// CHECK-32-EX-NEXT: [[TMP6:%.*]] = icmp eq i32 [[TMP5]], 1
-// CHECK-32-EX-NEXT: br i1 [[TMP6]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-32-EX-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK-32-EX-NEXT: store ptr [[E1]], ptr [[TMP2]], align 4
+// CHECK-32-EX-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
+// CHECK-32-EX-NEXT: [[TMP4:%.*]] = icmp eq i32 [[TMP3]], 1
+// CHECK-32-EX-NEXT: br i1 [[TMP4]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK-32-EX: .omp.reduction.then:
-// CHECK-32-EX-NEXT: [[TMP7:%.*]] = load double, ptr [[TMP0]], align 8
-// CHECK-32-EX-NEXT: [[TMP8:%.*]] = load double, ptr [[E1]], align 8
-// CHECK-32-EX-NEXT: [[ADD2:%.*]] = fadd double [[TMP7]], [[TMP8]]
+// CHECK-32-EX-NEXT: [[TMP5:%.*]] = load double, ptr [[TMP0]], align 8
+// CHECK-32-EX-NEXT: [[TMP6:%.*]] = load double, ptr [[E1]], align 8
+// CHECK-32-EX-NEXT: [[ADD2:%.*]] = fadd double [[TMP5]], [[TMP6]]
// CHECK-32-EX-NEXT: store double [[ADD2]], ptr [[TMP0]], align 8
// CHECK-32-EX-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK-32-EX: .omp.reduction.done:
@@ -1677,26 +1663,24 @@ int bar(int n){
// CHECK-32-EX-NEXT: [[TMP3:%.*]] = load float, ptr [[D2]], align 4
// CHECK-32-EX-NEXT: [[MUL:%.*]] = fmul float [[TMP3]], 3.300000e+01
// CHECK-32-EX-NEXT: store float [[MUL]], ptr [[D2]], align 4
-// CHECK-32-EX-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK-32-EX-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP4]], align 4
-// CHECK-32-EX-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK-32-EX-NEXT: store ptr [[C1]], ptr [[TMP6]], align 4
-// CHECK-32-EX-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
-// CHECK-32-EX-NEXT: store ptr [[D2]], ptr [[TMP7]], align 4
-// CHECK-32-EX-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP5]], i32 2, i32 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2)
-// CHECK-32-EX-NEXT: [[TMP9:%.*]] = icmp eq i32 [[TMP8]], 1
-// CHECK-32-EX-NEXT: br i1 [[TMP9]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-32-EX-NEXT: [[TMP4:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK-32-EX-NEXT: store ptr [[C1]], ptr [[TMP4]], align 4
+// CHECK-32-EX-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
+// CHECK-32-EX-NEXT: store ptr [[D2]], ptr [[TMP5]], align 4
+// CHECK-32-EX-NEXT: [[TMP6:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2)
+// CHECK-32-EX-NEXT: [[TMP7:%.*]] = icmp eq i32 [[TMP6]], 1
+// CHECK-32-EX-NEXT: br i1 [[TMP7]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK-32-EX: .omp.reduction.then:
-// CHECK-32-EX-NEXT: [[TMP10:%.*]] = load i8, ptr [[TMP0]], align 1
-// CHECK-32-EX-NEXT: [[CONV4:%.*]] = sext i8 [[TMP10]] to i32
-// CHECK-32-EX-NEXT: [[TMP11:%.*]] = load i8, ptr [[C1]], align 1
-// CHECK-32-EX-NEXT: [[CONV5:%.*]] = sext i8 [[TMP11]] to i32
+// CHECK-32-EX-NEXT: [[TMP8:%.*]] = load i8, ptr [[TMP0]], align 1
+// CHECK-32-EX-NEXT: [[CONV4:%.*]] = sext i8 [[TMP8]] to i32
+// CHECK-32-EX-NEXT: [[TMP9:%.*]] = load i8, ptr [[C1]], align 1
+// CHECK-32-EX-NEXT: [[CONV5:%.*]] = sext i8 [[TMP9]] to i32
// CHECK-32-EX-NEXT: [[XOR6:%.*]] = xor i32 [[CONV4]], [[CONV5]]
// CHECK-32-EX-NEXT: [[CONV7:%.*]] = trunc i32 [[XOR6]] to i8
// CHECK-32-EX-NEXT: store i8 [[CONV7]], ptr [[TMP0]], align 1
-// CHECK-32-EX-NEXT: [[TMP12:%.*]] = load float, ptr [[TMP1]], align 4
-// CHECK-32-EX-NEXT: [[TMP13:%.*]] = load float, ptr [[D2]], align 4
-// CHECK-32-EX-NEXT: [[MUL8:%.*]] = fmul float [[TMP12]], [[TMP13]]
+// CHECK-32-EX-NEXT: [[TMP10:%.*]] = load float, ptr [[TMP1]], align 4
+// CHECK-32-EX-NEXT: [[TMP11:%.*]] = load float, ptr [[D2]], align 4
+// CHECK-32-EX-NEXT: [[MUL8:%.*]] = fmul float [[TMP10]], [[TMP11]]
// CHECK-32-EX-NEXT: store float [[MUL8]], ptr [[TMP1]], align 4
// CHECK-32-EX-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK-32-EX: .omp.reduction.done:
@@ -1924,34 +1908,32 @@ int bar(int n){
// CHECK-32-EX-NEXT: [[COND:%.*]] = phi i32 [ 99, [[COND_TRUE]] ], [ [[CONV3]], [[COND_FALSE]] ]
// CHECK-32-EX-NEXT: [[CONV4:%.*]] = trunc i32 [[COND]] to i16
// CHECK-32-EX-NEXT: store i16 [[CONV4]], ptr [[B2]], align 2
-// CHECK-32-EX-NEXT: [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK-32-EX-NEXT: [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
-// CHECK-32-EX-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK-32-EX-NEXT: store ptr [[A1]], ptr [[TMP7]], align 4
-// CHECK-32-EX-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
-// CHECK-32-EX-NEXT: store ptr [[B2]], ptr [[TMP8]], align 4
-// CHECK-32-EX-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP6]], i32 2, i32 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func3, ptr @_omp_reduction_inter_warp_copy_func4)
-// CHECK-32-EX-NEXT: [[TMP10:%.*]] = icmp eq i32 [[TMP9]], 1
-// CHECK-32-EX-NEXT: br i1 [[TMP10]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-32-EX-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK-32-EX-NEXT: store ptr [[A1]], ptr [[TMP5]], align 4
+// CHECK-32-EX-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
+// CHECK-32-EX-NEXT: store ptr [[B2]], ptr [[TMP6]], align 4
+// CHECK-32-EX-NEXT: [[TMP7:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func3, ptr @_omp_reduction_inter_warp_copy_func4)
+// CHECK-32-EX-NEXT: [[TMP8:%.*]] = icmp eq i32 [[TMP7]], 1
+// CHECK-32-EX-NEXT: br i1 [[TMP8]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK-32-EX: .omp.reduction.then:
-// CHECK-32-EX-NEXT: [[TMP11:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK-32-EX-NEXT: [[TMP12:%.*]] = load i32, ptr [[A1]], align 4
-// CHECK-32-EX-NEXT: [[OR5:%.*]] = or i32 [[TMP11]], [[TMP12]]
+// CHECK-32-EX-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK-32-EX-NEXT: [[TMP10:%.*]] = load i32, ptr [[A1]], align 4
+// CHECK-32-EX-NEXT: [[OR5:%.*]] = or i32 [[TMP9]], [[TMP10]]
// CHECK-32-EX-NEXT: store i32 [[OR5]], ptr [[TMP0]], align 4
-// CHECK-32-EX-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
-// CHECK-32-EX-NEXT: [[CONV6:%.*]] = sext i16 [[TMP13]] to i32
-// CHECK-32-EX-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
-// CHECK-32-EX-NEXT: [[CONV7:%.*]] = sext i16 [[TMP14]] to i32
+// CHECK-32-EX-NEXT: [[TMP11:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK-32-EX-NEXT: [[CONV6:%.*]] = sext i16 [[TMP11]] to i32
+// CHECK-32-EX-NEXT: [[TMP12:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK-32-EX-NEXT: [[CONV7:%.*]] = sext i16 [[TMP12]] to i32
// CHECK-32-EX-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[CONV6]], [[CONV7]]
// CHECK-32-EX-NEXT: br i1 [[CMP8]], label [[COND_TRUE9:%.*]], label [[COND_FALSE10:%.*]]
// CHECK-32-EX: cond.true9:
-// CHECK-32-EX-NEXT: [[TMP15:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK-32-EX-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
// CHECK-32-EX-NEXT: br label [[COND_END11:%.*]]
// CHECK-32-EX: cond.false10:
-// CHECK-32-EX-NEXT: [[TMP16:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK-32-EX-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
// CHECK-32-EX-NEXT: br label [[COND_END11]]
// CHECK-32-EX: cond.end11:
-// CHECK-32-EX-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP15]], [[COND_TRUE9]] ], [ [[TMP16]], [[COND_FALSE10]] ]
+// CHECK-32-EX-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP13]], [[COND_TRUE9]] ], [ [[TMP14]], [[COND_FALSE10]] ]
// CHECK-32-EX-NEXT: store i16 [[COND12]], ptr [[TMP1]], align 2
// CHECK-32-EX-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK-32-EX: .omp.reduction.done:
diff --git a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp
index d4bb6c1639e4609..fac42732022ad85 100644
--- a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp
+++ b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp
@@ -329,13 +329,11 @@ void test() {
// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP32:%.*]] = load i32, ptr [[TMP31]], align 4, !tbaa [[TBAA14]]
// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP32]])
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = load i32, ptr [[TMP33]], align 4, !tbaa [[TBAA14]]
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[PARTIAL_SUM5]], ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP34]], i32 1, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
-// CHECK1-NEXT: [[TMP37:%.*]] = icmp eq i32 [[TMP36]], 1
-// CHECK1-NEXT: br i1 [[TMP37]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[PARTIAL_SUM5]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
+// CHECK1-NEXT: [[TMP35:%.*]] = icmp eq i32 [[TMP34]], 1
+// CHECK1-NEXT: br i1 [[TMP35]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK1: .omp.reduction.then:
// CHECK1-NEXT: [[CALL21:%.*]] = call nonnull align 4 dereferenceable(8) ptr @_ZNSt7complexIfEpLIfEERS0_RKS_IT_E(ptr nonnull align 4 dereferenceable(8) [[TMP2]], ptr nonnull align 4 dereferenceable(8) [[PARTIAL_SUM5]]) #[[ATTR12]]
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
@@ -822,13 +820,11 @@ void test() {
// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP32:%.*]] = load i32, ptr [[TMP31]], align 4, !tbaa [[TBAA14]]
// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP32]])
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = load i32, ptr [[TMP33]], align 4, !tbaa [[TBAA14]]
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[PARTIAL_SUM5]], ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP34]], i32 1, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2)
-// CHECK1-NEXT: [[TMP37:%.*]] = icmp eq i32 [[TMP36]], 1
-// CHECK1-NEXT: br i1 [[TMP37]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[PARTIAL_SUM5]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2)
+// CHECK1-NEXT: [[TMP35:%.*]] = icmp eq i32 [[TMP34]], 1
+// CHECK1-NEXT: br i1 [[TMP35]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK1: .omp.reduction.then:
// CHECK1-NEXT: [[CALL21:%.*]] = call nonnull align 8 dereferenceable(16) ptr @_ZNSt7complexIdEpLIdEERS0_RKS_IT_E(ptr nonnull align 8 dereferenceable(16) [[TMP2]], ptr nonnull align 8 dereferenceable(16) [[PARTIAL_SUM5]]) #[[ATTR12]]
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
diff --git a/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp b/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp
index 27af206098c10b1..ef4a695975d9519 100644
--- a/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp
@@ -93,18 +93,16 @@ int bar(int n){
// CHECK1-NEXT: [[TMP1:%.*]] = load double, ptr [[E1]], align 8
// CHECK1-NEXT: [[ADD:%.*]] = fadd double [[TMP1]], 5.000000e+00
// CHECK1-NEXT: store double [[ADD]], ptr [[E1]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
-// CHECK1-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[E1]], ptr [[TMP4]], align 8
+// CHECK1-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[E1]], ptr [[TMP2]], align 8
// CHECK1-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// CHECK1-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP3]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func, ptr @_omp_reduction_list_to_global_copy_func, ptr @_omp_reduction_list_to_global_reduce_func, ptr @_omp_reduction_global_to_list_copy_func, ptr @_omp_reduction_global_to_list_reduce_func)
-// CHECK1-NEXT: [[TMP6:%.*]] = icmp eq i32 [[TMP5]], 1
-// CHECK1-NEXT: br i1 [[TMP6]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK1-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func, ptr @_omp_reduction_list_to_global_copy_func, ptr @_omp_reduction_list_to_global_reduce_func, ptr @_omp_reduction_global_to_list_copy_func, ptr @_omp_reduction_global_to_list_reduce_func)
+// CHECK1-NEXT: [[TMP4:%.*]] = icmp eq i32 [[TMP3]], 1
+// CHECK1-NEXT: br i1 [[TMP4]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK1: .omp.reduction.then:
-// CHECK1-NEXT: [[TMP7:%.*]] = load double, ptr [[TMP0]], align 8
-// CHECK1-NEXT: [[TMP8:%.*]] = load double, ptr [[E1]], align 8
-// CHECK1-NEXT: [[ADD2:%.*]] = fadd double [[TMP7]], [[TMP8]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load double, ptr [[TMP0]], align 8
+// CHECK1-NEXT: [[TMP6:%.*]] = load double, ptr [[E1]], align 8
+// CHECK1-NEXT: [[ADD2:%.*]] = fadd double [[TMP5]], [[TMP6]]
// CHECK1-NEXT: store double [[ADD2]], ptr [[TMP0]], align 8
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK1: .omp.reduction.done:
@@ -251,7 +249,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i64 0, i64 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x double], ptr [[E]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP9:%.*]] = load double, ptr [[TMP7]], align 8
// CHECK1-NEXT: store double [[TMP9]], ptr [[TMP8]], align 8
// CHECK1-NEXT: ret void
@@ -271,7 +269,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x double], ptr [[E]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 8
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTADDR2]], align 8
// CHECK1-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l20_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr [[TMP7]]) #[[ATTR4]]
@@ -293,7 +291,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i64 0, i64 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x double], ptr [[E]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP9:%.*]] = load double, ptr [[TMP8]], align 8
// CHECK1-NEXT: store double [[TMP9]], ptr [[TMP7]], align 8
// CHECK1-NEXT: ret void
@@ -313,7 +311,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x double], ptr [[E]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 8
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTADDR2]], align 8
// CHECK1-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l20_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP7]], ptr [[DOTOMP_REDUCTION_RED_LIST]]) #[[ATTR4]]
@@ -379,27 +377,25 @@ int bar(int n){
// CHECK1-NEXT: [[TMP3:%.*]] = load float, ptr [[D2]], align 4
// CHECK1-NEXT: [[MUL:%.*]] = fmul float [[TMP3]], 3.300000e+01
// CHECK1-NEXT: store float [[MUL]], ptr [[D2]], align 4
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP4]], align 4
-// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[C1]], ptr [[TMP6]], align 8
-// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[D2]], ptr [[TMP7]], align 8
+// CHECK1-NEXT: [[TMP4:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[C1]], ptr [[TMP4]], align 8
+// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[D2]], ptr [[TMP5]], align 8
// CHECK1-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP5]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2, ptr @_omp_reduction_list_to_global_copy_func3, ptr @_omp_reduction_list_to_global_reduce_func4, ptr @_omp_reduction_global_to_list_copy_func5, ptr @_omp_reduction_global_to_list_reduce_func6)
-// CHECK1-NEXT: [[TMP9:%.*]] = icmp eq i32 [[TMP8]], 1
-// CHECK1-NEXT: br i1 [[TMP9]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK1-NEXT: [[TMP6:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2, ptr @_omp_reduction_list_to_global_copy_func3, ptr @_omp_reduction_list_to_global_reduce_func4, ptr @_omp_reduction_global_to_list_copy_func5, ptr @_omp_reduction_global_to_list_reduce_func6)
+// CHECK1-NEXT: [[TMP7:%.*]] = icmp eq i32 [[TMP6]], 1
+// CHECK1-NEXT: br i1 [[TMP7]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK1: .omp.reduction.then:
-// CHECK1-NEXT: [[TMP10:%.*]] = load i8, ptr [[TMP0]], align 1
-// CHECK1-NEXT: [[CONV4:%.*]] = sext i8 [[TMP10]] to i32
-// CHECK1-NEXT: [[TMP11:%.*]] = load i8, ptr [[C1]], align 1
-// CHECK1-NEXT: [[CONV5:%.*]] = sext i8 [[TMP11]] to i32
+// CHECK1-NEXT: [[TMP8:%.*]] = load i8, ptr [[TMP0]], align 1
+// CHECK1-NEXT: [[CONV4:%.*]] = sext i8 [[TMP8]] to i32
+// CHECK1-NEXT: [[TMP9:%.*]] = load i8, ptr [[C1]], align 1
+// CHECK1-NEXT: [[CONV5:%.*]] = sext i8 [[TMP9]] to i32
// CHECK1-NEXT: [[XOR6:%.*]] = xor i32 [[CONV4]], [[CONV5]]
// CHECK1-NEXT: [[CONV7:%.*]] = trunc i32 [[XOR6]] to i8
// CHECK1-NEXT: store i8 [[CONV7]], ptr [[TMP0]], align 1
-// CHECK1-NEXT: [[TMP12:%.*]] = load float, ptr [[TMP1]], align 4
-// CHECK1-NEXT: [[TMP13:%.*]] = load float, ptr [[D2]], align 4
-// CHECK1-NEXT: [[MUL8:%.*]] = fmul float [[TMP12]], [[TMP13]]
+// CHECK1-NEXT: [[TMP10:%.*]] = load float, ptr [[TMP1]], align 4
+// CHECK1-NEXT: [[TMP11:%.*]] = load float, ptr [[D2]], align 4
+// CHECK1-NEXT: [[MUL8:%.*]] = fmul float [[TMP10]], [[TMP11]]
// CHECK1-NEXT: store float [[MUL8]], ptr [[TMP1]], align 4
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK1: .omp.reduction.done:
@@ -581,13 +577,13 @@ int bar(int n){
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i64 0, i64 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP9:%.*]] = load i8, ptr [[TMP7]], align 1
// CHECK1-NEXT: store i8 [[TMP9]], ptr [[TMP8]], align 4
// CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i64 0, i64 1
// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 8
// CHECK1-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP4]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1024 x float], ptr [[D]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP13:%.*]] = load float, ptr [[TMP11]], align 4
// CHECK1-NEXT: store float [[TMP13]], ptr [[TMP12]], align 4
// CHECK1-NEXT: ret void
@@ -607,11 +603,11 @@ int bar(int n){
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 8
// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
// CHECK1-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x float], ptr [[D]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 8
// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 8
// CHECK1-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l26_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr [[TMP9]]) #[[ATTR4]]
@@ -633,13 +629,13 @@ int bar(int n){
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i64 0, i64 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP9:%.*]] = load i8, ptr [[TMP8]], align 4
// CHECK1-NEXT: store i8 [[TMP9]], ptr [[TMP7]], align 1
// CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i64 0, i64 1
// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 8
// CHECK1-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP4]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1024 x float], ptr [[D]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP13:%.*]] = load float, ptr [[TMP12]], align 4
// CHECK1-NEXT: store float [[TMP13]], ptr [[TMP11]], align 4
// CHECK1-NEXT: ret void
@@ -659,11 +655,11 @@ int bar(int n){
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 8
// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
// CHECK1-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x float], ptr [[D]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 8
// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 8
// CHECK1-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l26_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP9]], ptr [[DOTOMP_REDUCTION_RED_LIST]]) #[[ATTR4]]
@@ -726,7 +722,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
// CHECK1-NEXT: store ptr [[B2]], ptr [[TMP7]], align 8
// CHECK1-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP5]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func9, ptr @_omp_reduction_inter_warp_copy_func10, ptr @_omp_reduction_list_to_global_copy_func11, ptr @_omp_reduction_list_to_global_reduce_func12, ptr @_omp_reduction_global_to_list_copy_func13, ptr @_omp_reduction_global_to_list_reduce_func14)
+// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func9, ptr @_omp_reduction_inter_warp_copy_func10, ptr @_omp_reduction_list_to_global_copy_func11, ptr @_omp_reduction_list_to_global_reduce_func12, ptr @_omp_reduction_global_to_list_copy_func13, ptr @_omp_reduction_global_to_list_reduce_func14)
// CHECK1-NEXT: [[TMP9:%.*]] = icmp eq i32 [[TMP8]], 1
// CHECK1-NEXT: br i1 [[TMP9]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK1: .omp.reduction.then:
@@ -789,34 +785,32 @@ int bar(int n){
// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 99, [[COND_TRUE]] ], [ [[CONV3]], [[COND_FALSE]] ]
// CHECK1-NEXT: [[CONV4:%.*]] = trunc i32 [[COND]] to i16
// CHECK1-NEXT: store i16 [[CONV4]], ptr [[B2]], align 2
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
-// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[A1]], ptr [[TMP7]], align 8
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[B2]], ptr [[TMP8]], align 8
-// CHECK1-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP6]], i32 2, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func7, ptr @_omp_reduction_inter_warp_copy_func8)
-// CHECK1-NEXT: [[TMP10:%.*]] = icmp eq i32 [[TMP9]], 1
-// CHECK1-NEXT: br i1 [[TMP10]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[A1]], ptr [[TMP5]], align 8
+// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[B2]], ptr [[TMP6]], align 8
+// CHECK1-NEXT: [[TMP7:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func7, ptr @_omp_reduction_inter_warp_copy_func8)
+// CHECK1-NEXT: [[TMP8:%.*]] = icmp eq i32 [[TMP7]], 1
+// CHECK1-NEXT: br i1 [[TMP8]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK1: .omp.reduction.then:
-// CHECK1-NEXT: [[TMP11:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP12:%.*]] = load i32, ptr [[A1]], align 4
-// CHECK1-NEXT: [[OR5:%.*]] = or i32 [[TMP11]], [[TMP12]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP10:%.*]] = load i32, ptr [[A1]], align 4
+// CHECK1-NEXT: [[OR5:%.*]] = or i32 [[TMP9]], [[TMP10]]
// CHECK1-NEXT: store i32 [[OR5]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
-// CHECK1-NEXT: [[CONV6:%.*]] = sext i16 [[TMP13]] to i32
-// CHECK1-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
-// CHECK1-NEXT: [[CONV7:%.*]] = sext i16 [[TMP14]] to i32
+// CHECK1-NEXT: [[TMP11:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK1-NEXT: [[CONV6:%.*]] = sext i16 [[TMP11]] to i32
+// CHECK1-NEXT: [[TMP12:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK1-NEXT: [[CONV7:%.*]] = sext i16 [[TMP12]] to i32
// CHECK1-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[CONV6]], [[CONV7]]
// CHECK1-NEXT: br i1 [[CMP8]], label [[COND_TRUE9:%.*]], label [[COND_FALSE10:%.*]]
// CHECK1: cond.true9:
-// CHECK1-NEXT: [[TMP15:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK1-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
// CHECK1-NEXT: br label [[COND_END11:%.*]]
// CHECK1: cond.false10:
-// CHECK1-NEXT: [[TMP16:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK1-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
// CHECK1-NEXT: br label [[COND_END11]]
// CHECK1: cond.end11:
-// CHECK1-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP15]], [[COND_TRUE9]] ], [ [[TMP16]], [[COND_FALSE10]] ]
+// CHECK1-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP13]], [[COND_TRUE9]] ], [ [[TMP14]], [[COND_FALSE10]] ]
// CHECK1-NEXT: store i16 [[COND12]], ptr [[TMP1]], align 2
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK1: .omp.reduction.done:
@@ -1154,13 +1148,13 @@ int bar(int n){
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i64 0, i64 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK1-NEXT: store i32 [[TMP9]], ptr [[TMP8]], align 4
// CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i64 0, i64 1
// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 8
// CHECK1-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP4]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1024 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP11]], align 2
// CHECK1-NEXT: store i16 [[TMP13]], ptr [[TMP12]], align 4
// CHECK1-NEXT: ret void
@@ -1180,11 +1174,11 @@ int bar(int n){
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 8
// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
// CHECK1-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 8
// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 8
// CHECK1-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l33_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr [[TMP9]]) #[[ATTR4]]
@@ -1206,13 +1200,13 @@ int bar(int n){
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i64 0, i64 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP8]], align 4
// CHECK1-NEXT: store i32 [[TMP9]], ptr [[TMP7]], align 4
// CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i64 0, i64 1
// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 8
// CHECK1-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP4]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1024 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
// CHECK1-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP12]], align 4
// CHECK1-NEXT: store i16 [[TMP13]], ptr [[TMP11]], align 2
// CHECK1-NEXT: ret void
@@ -1232,11 +1226,11 @@ int bar(int n){
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 8
// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
// CHECK1-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
// CHECK1-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 8
// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 8
// CHECK1-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l33_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP9]], ptr [[DOTOMP_REDUCTION_RED_LIST]]) #[[ATTR4]]
@@ -1286,18 +1280,16 @@ int bar(int n){
// CHECK2-NEXT: [[TMP1:%.*]] = load double, ptr [[E1]], align 8
// CHECK2-NEXT: [[ADD:%.*]] = fadd double [[TMP1]], 5.000000e+00
// CHECK2-NEXT: store double [[ADD]], ptr [[E1]], align 8
-// CHECK2-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK2-NEXT: [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
-// CHECK2-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK2-NEXT: store ptr [[E1]], ptr [[TMP4]], align 4
+// CHECK2-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK2-NEXT: store ptr [[E1]], ptr [[TMP2]], align 4
// CHECK2-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// CHECK2-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP3]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func, ptr @_omp_reduction_list_to_global_copy_func, ptr @_omp_reduction_list_to_global_reduce_func, ptr @_omp_reduction_global_to_list_copy_func, ptr @_omp_reduction_global_to_list_reduce_func)
-// CHECK2-NEXT: [[TMP6:%.*]] = icmp eq i32 [[TMP5]], 1
-// CHECK2-NEXT: br i1 [[TMP6]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK2-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func, ptr @_omp_reduction_list_to_global_copy_func, ptr @_omp_reduction_list_to_global_reduce_func, ptr @_omp_reduction_global_to_list_copy_func, ptr @_omp_reduction_global_to_list_reduce_func)
+// CHECK2-NEXT: [[TMP4:%.*]] = icmp eq i32 [[TMP3]], 1
+// CHECK2-NEXT: br i1 [[TMP4]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK2: .omp.reduction.then:
-// CHECK2-NEXT: [[TMP7:%.*]] = load double, ptr [[TMP0]], align 8
-// CHECK2-NEXT: [[TMP8:%.*]] = load double, ptr [[E1]], align 8
-// CHECK2-NEXT: [[ADD2:%.*]] = fadd double [[TMP7]], [[TMP8]]
+// CHECK2-NEXT: [[TMP5:%.*]] = load double, ptr [[TMP0]], align 8
+// CHECK2-NEXT: [[TMP6:%.*]] = load double, ptr [[E1]], align 8
+// CHECK2-NEXT: [[ADD2:%.*]] = fadd double [[TMP5]], [[TMP6]]
// CHECK2-NEXT: store double [[ADD2]], ptr [[TMP0]], align 8
// CHECK2-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK2: .omp.reduction.done:
@@ -1444,7 +1436,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK2-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK2-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x double], ptr [[E]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP9:%.*]] = load double, ptr [[TMP7]], align 8
// CHECK2-NEXT: store double [[TMP9]], ptr [[TMP8]], align 8
// CHECK2-NEXT: ret void
@@ -1464,7 +1456,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK2-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x double], ptr [[E]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK2-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK2-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l20_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr [[TMP7]]) #[[ATTR4]]
@@ -1486,7 +1478,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK2-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK2-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x double], ptr [[E]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP9:%.*]] = load double, ptr [[TMP8]], align 8
// CHECK2-NEXT: store double [[TMP9]], ptr [[TMP7]], align 8
// CHECK2-NEXT: ret void
@@ -1506,7 +1498,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK2-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x double], ptr [[E]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK2-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK2-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l20_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP7]], ptr [[DOTOMP_REDUCTION_RED_LIST]]) #[[ATTR4]]
@@ -1572,27 +1564,25 @@ int bar(int n){
// CHECK2-NEXT: [[TMP3:%.*]] = load float, ptr [[D2]], align 4
// CHECK2-NEXT: [[MUL:%.*]] = fmul float [[TMP3]], 3.300000e+01
// CHECK2-NEXT: store float [[MUL]], ptr [[D2]], align 4
-// CHECK2-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK2-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP4]], align 4
-// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK2-NEXT: store ptr [[C1]], ptr [[TMP6]], align 4
-// CHECK2-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
-// CHECK2-NEXT: store ptr [[D2]], ptr [[TMP7]], align 4
+// CHECK2-NEXT: [[TMP4:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK2-NEXT: store ptr [[C1]], ptr [[TMP4]], align 4
+// CHECK2-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
+// CHECK2-NEXT: store ptr [[D2]], ptr [[TMP5]], align 4
// CHECK2-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// CHECK2-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP5]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2, ptr @_omp_reduction_list_to_global_copy_func3, ptr @_omp_reduction_list_to_global_reduce_func4, ptr @_omp_reduction_global_to_list_copy_func5, ptr @_omp_reduction_global_to_list_reduce_func6)
-// CHECK2-NEXT: [[TMP9:%.*]] = icmp eq i32 [[TMP8]], 1
-// CHECK2-NEXT: br i1 [[TMP9]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK2-NEXT: [[TMP6:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2, ptr @_omp_reduction_list_to_global_copy_func3, ptr @_omp_reduction_list_to_global_reduce_func4, ptr @_omp_reduction_global_to_list_copy_func5, ptr @_omp_reduction_global_to_list_reduce_func6)
+// CHECK2-NEXT: [[TMP7:%.*]] = icmp eq i32 [[TMP6]], 1
+// CHECK2-NEXT: br i1 [[TMP7]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK2: .omp.reduction.then:
-// CHECK2-NEXT: [[TMP10:%.*]] = load i8, ptr [[TMP0]], align 1
-// CHECK2-NEXT: [[CONV4:%.*]] = sext i8 [[TMP10]] to i32
-// CHECK2-NEXT: [[TMP11:%.*]] = load i8, ptr [[C1]], align 1
-// CHECK2-NEXT: [[CONV5:%.*]] = sext i8 [[TMP11]] to i32
+// CHECK2-NEXT: [[TMP8:%.*]] = load i8, ptr [[TMP0]], align 1
+// CHECK2-NEXT: [[CONV4:%.*]] = sext i8 [[TMP8]] to i32
+// CHECK2-NEXT: [[TMP9:%.*]] = load i8, ptr [[C1]], align 1
+// CHECK2-NEXT: [[CONV5:%.*]] = sext i8 [[TMP9]] to i32
// CHECK2-NEXT: [[XOR6:%.*]] = xor i32 [[CONV4]], [[CONV5]]
// CHECK2-NEXT: [[CONV7:%.*]] = trunc i32 [[XOR6]] to i8
// CHECK2-NEXT: store i8 [[CONV7]], ptr [[TMP0]], align 1
-// CHECK2-NEXT: [[TMP12:%.*]] = load float, ptr [[TMP1]], align 4
-// CHECK2-NEXT: [[TMP13:%.*]] = load float, ptr [[D2]], align 4
-// CHECK2-NEXT: [[MUL8:%.*]] = fmul float [[TMP12]], [[TMP13]]
+// CHECK2-NEXT: [[TMP10:%.*]] = load float, ptr [[TMP1]], align 4
+// CHECK2-NEXT: [[TMP11:%.*]] = load float, ptr [[D2]], align 4
+// CHECK2-NEXT: [[MUL8:%.*]] = fmul float [[TMP10]], [[TMP11]]
// CHECK2-NEXT: store float [[MUL8]], ptr [[TMP1]], align 4
// CHECK2-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK2: .omp.reduction.done:
@@ -1774,13 +1764,13 @@ int bar(int n){
// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK2-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK2-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP9:%.*]] = load i8, ptr [[TMP7]], align 1
// CHECK2-NEXT: store i8 [[TMP9]], ptr [[TMP8]], align 4
// CHECK2-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 1
// CHECK2-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 4
// CHECK2-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP4]], i32 0, i32 1
-// CHECK2-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1024 x float], ptr [[D]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP13:%.*]] = load float, ptr [[TMP11]], align 4
// CHECK2-NEXT: store float [[TMP13]], ptr [[TMP12]], align 4
// CHECK2-NEXT: ret void
@@ -1800,11 +1790,11 @@ int bar(int n){
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK2-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK2-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK2-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP3]], i32 0, i32 1
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x float], ptr [[D]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 4
// CHECK2-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK2-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l26_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr [[TMP9]]) #[[ATTR4]]
@@ -1826,13 +1816,13 @@ int bar(int n){
// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK2-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK2-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP9:%.*]] = load i8, ptr [[TMP8]], align 4
// CHECK2-NEXT: store i8 [[TMP9]], ptr [[TMP7]], align 1
// CHECK2-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 1
// CHECK2-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 4
// CHECK2-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP4]], i32 0, i32 1
-// CHECK2-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1024 x float], ptr [[D]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP13:%.*]] = load float, ptr [[TMP12]], align 4
// CHECK2-NEXT: store float [[TMP13]], ptr [[TMP11]], align 4
// CHECK2-NEXT: ret void
@@ -1852,11 +1842,11 @@ int bar(int n){
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK2-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK2-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK2-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP3]], i32 0, i32 1
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x float], ptr [[D]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 4
// CHECK2-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK2-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l26_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP9]], ptr [[DOTOMP_REDUCTION_RED_LIST]]) #[[ATTR4]]
@@ -1919,7 +1909,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK2-NEXT: store ptr [[B2]], ptr [[TMP7]], align 4
// CHECK2-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// CHECK2-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP5]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func9, ptr @_omp_reduction_inter_warp_copy_func10, ptr @_omp_reduction_list_to_global_copy_func11, ptr @_omp_reduction_list_to_global_reduce_func12, ptr @_omp_reduction_global_to_list_copy_func13, ptr @_omp_reduction_global_to_list_reduce_func14)
+// CHECK2-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func9, ptr @_omp_reduction_inter_warp_copy_func10, ptr @_omp_reduction_list_to_global_copy_func11, ptr @_omp_reduction_list_to_global_reduce_func12, ptr @_omp_reduction_global_to_list_copy_func13, ptr @_omp_reduction_global_to_list_reduce_func14)
// CHECK2-NEXT: [[TMP9:%.*]] = icmp eq i32 [[TMP8]], 1
// CHECK2-NEXT: br i1 [[TMP9]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK2: .omp.reduction.then:
@@ -1982,34 +1972,32 @@ int bar(int n){
// CHECK2-NEXT: [[COND:%.*]] = phi i32 [ 99, [[COND_TRUE]] ], [ [[CONV3]], [[COND_FALSE]] ]
// CHECK2-NEXT: [[CONV4:%.*]] = trunc i32 [[COND]] to i16
// CHECK2-NEXT: store i16 [[CONV4]], ptr [[B2]], align 2
-// CHECK2-NEXT: [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK2-NEXT: [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
-// CHECK2-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK2-NEXT: store ptr [[A1]], ptr [[TMP7]], align 4
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
-// CHECK2-NEXT: store ptr [[B2]], ptr [[TMP8]], align 4
-// CHECK2-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP6]], i32 2, i32 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func7, ptr @_omp_reduction_inter_warp_copy_func8)
-// CHECK2-NEXT: [[TMP10:%.*]] = icmp eq i32 [[TMP9]], 1
-// CHECK2-NEXT: br i1 [[TMP10]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK2-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK2-NEXT: store ptr [[A1]], ptr [[TMP5]], align 4
+// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
+// CHECK2-NEXT: store ptr [[B2]], ptr [[TMP6]], align 4
+// CHECK2-NEXT: [[TMP7:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func7, ptr @_omp_reduction_inter_warp_copy_func8)
+// CHECK2-NEXT: [[TMP8:%.*]] = icmp eq i32 [[TMP7]], 1
+// CHECK2-NEXT: br i1 [[TMP8]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK2: .omp.reduction.then:
-// CHECK2-NEXT: [[TMP11:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK2-NEXT: [[TMP12:%.*]] = load i32, ptr [[A1]], align 4
-// CHECK2-NEXT: [[OR5:%.*]] = or i32 [[TMP11]], [[TMP12]]
+// CHECK2-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK2-NEXT: [[TMP10:%.*]] = load i32, ptr [[A1]], align 4
+// CHECK2-NEXT: [[OR5:%.*]] = or i32 [[TMP9]], [[TMP10]]
// CHECK2-NEXT: store i32 [[OR5]], ptr [[TMP0]], align 4
-// CHECK2-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
-// CHECK2-NEXT: [[CONV6:%.*]] = sext i16 [[TMP13]] to i32
-// CHECK2-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
-// CHECK2-NEXT: [[CONV7:%.*]] = sext i16 [[TMP14]] to i32
+// CHECK2-NEXT: [[TMP11:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK2-NEXT: [[CONV6:%.*]] = sext i16 [[TMP11]] to i32
+// CHECK2-NEXT: [[TMP12:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK2-NEXT: [[CONV7:%.*]] = sext i16 [[TMP12]] to i32
// CHECK2-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[CONV6]], [[CONV7]]
// CHECK2-NEXT: br i1 [[CMP8]], label [[COND_TRUE9:%.*]], label [[COND_FALSE10:%.*]]
// CHECK2: cond.true9:
-// CHECK2-NEXT: [[TMP15:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK2-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
// CHECK2-NEXT: br label [[COND_END11:%.*]]
// CHECK2: cond.false10:
-// CHECK2-NEXT: [[TMP16:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK2-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
// CHECK2-NEXT: br label [[COND_END11]]
// CHECK2: cond.end11:
-// CHECK2-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP15]], [[COND_TRUE9]] ], [ [[TMP16]], [[COND_FALSE10]] ]
+// CHECK2-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP13]], [[COND_TRUE9]] ], [ [[TMP14]], [[COND_FALSE10]] ]
// CHECK2-NEXT: store i16 [[COND12]], ptr [[TMP1]], align 2
// CHECK2-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK2: .omp.reduction.done:
@@ -2347,13 +2335,13 @@ int bar(int n){
// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK2-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK2-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK2-NEXT: store i32 [[TMP9]], ptr [[TMP8]], align 4
// CHECK2-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 1
// CHECK2-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 4
// CHECK2-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP4]], i32 0, i32 1
-// CHECK2-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1024 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP11]], align 2
// CHECK2-NEXT: store i16 [[TMP13]], ptr [[TMP12]], align 4
// CHECK2-NEXT: ret void
@@ -2373,11 +2361,11 @@ int bar(int n){
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK2-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK2-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK2-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP3]], i32 0, i32 1
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 4
// CHECK2-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK2-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l33_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr [[TMP9]]) #[[ATTR4]]
@@ -2399,13 +2387,13 @@ int bar(int n){
// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK2-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK2-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP8]], align 4
// CHECK2-NEXT: store i32 [[TMP9]], ptr [[TMP7]], align 4
// CHECK2-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 1
// CHECK2-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 4
// CHECK2-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP4]], i32 0, i32 1
-// CHECK2-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1024 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
+// CHECK2-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
// CHECK2-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP12]], align 4
// CHECK2-NEXT: store i16 [[TMP13]], ptr [[TMP11]], align 2
// CHECK2-NEXT: ret void
@@ -2425,11 +2413,11 @@ int bar(int n){
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK2-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK2-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK2-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP3]], i32 0, i32 1
-// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
+// CHECK2-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
// CHECK2-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 4
// CHECK2-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK2-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l33_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP9]], ptr [[DOTOMP_REDUCTION_RED_LIST]]) #[[ATTR4]]
@@ -2479,18 +2467,16 @@ int bar(int n){
// CHECK3-NEXT: [[TMP1:%.*]] = load double, ptr [[E1]], align 8
// CHECK3-NEXT: [[ADD:%.*]] = fadd double [[TMP1]], 5.000000e+00
// CHECK3-NEXT: store double [[ADD]], ptr [[E1]], align 8
-// CHECK3-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
-// CHECK3-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK3-NEXT: store ptr [[E1]], ptr [[TMP4]], align 4
+// CHECK3-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK3-NEXT: store ptr [[E1]], ptr [[TMP2]], align 4
// CHECK3-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// CHECK3-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP3]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 2048, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func, ptr @_omp_reduction_list_to_global_copy_func, ptr @_omp_reduction_list_to_global_reduce_func, ptr @_omp_reduction_global_to_list_copy_func, ptr @_omp_reduction_global_to_list_reduce_func)
-// CHECK3-NEXT: [[TMP6:%.*]] = icmp eq i32 [[TMP5]], 1
-// CHECK3-NEXT: br i1 [[TMP6]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK3-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 2048, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func, ptr @_omp_reduction_list_to_global_copy_func, ptr @_omp_reduction_list_to_global_reduce_func, ptr @_omp_reduction_global_to_list_copy_func, ptr @_omp_reduction_global_to_list_reduce_func)
+// CHECK3-NEXT: [[TMP4:%.*]] = icmp eq i32 [[TMP3]], 1
+// CHECK3-NEXT: br i1 [[TMP4]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK3: .omp.reduction.then:
-// CHECK3-NEXT: [[TMP7:%.*]] = load double, ptr [[TMP0]], align 8
-// CHECK3-NEXT: [[TMP8:%.*]] = load double, ptr [[E1]], align 8
-// CHECK3-NEXT: [[ADD2:%.*]] = fadd double [[TMP7]], [[TMP8]]
+// CHECK3-NEXT: [[TMP5:%.*]] = load double, ptr [[TMP0]], align 8
+// CHECK3-NEXT: [[TMP6:%.*]] = load double, ptr [[E1]], align 8
+// CHECK3-NEXT: [[ADD2:%.*]] = fadd double [[TMP5]], [[TMP6]]
// CHECK3-NEXT: store double [[ADD2]], ptr [[TMP0]], align 8
// CHECK3-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK3: .omp.reduction.done:
@@ -2637,7 +2623,7 @@ int bar(int n){
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x double], ptr [[E]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP9:%.*]] = load double, ptr [[TMP7]], align 8
// CHECK3-NEXT: store double [[TMP9]], ptr [[TMP8]], align 8
// CHECK3-NEXT: ret void
@@ -2657,7 +2643,7 @@ int bar(int n){
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK3-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2048 x double], ptr [[E]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK3-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l20_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr [[TMP7]]) #[[ATTR4]]
@@ -2679,7 +2665,7 @@ int bar(int n){
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x double], ptr [[E]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP9:%.*]] = load double, ptr [[TMP8]], align 8
// CHECK3-NEXT: store double [[TMP9]], ptr [[TMP7]], align 8
// CHECK3-NEXT: ret void
@@ -2699,7 +2685,7 @@ int bar(int n){
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK3-NEXT: [[E:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2048 x double], ptr [[E]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x double], ptr [[E]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK3-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l20_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP7]], ptr [[DOTOMP_REDUCTION_RED_LIST]]) #[[ATTR4]]
@@ -2765,27 +2751,25 @@ int bar(int n){
// CHECK3-NEXT: [[TMP3:%.*]] = load float, ptr [[D2]], align 4
// CHECK3-NEXT: [[MUL:%.*]] = fmul float [[TMP3]], 3.300000e+01
// CHECK3-NEXT: store float [[MUL]], ptr [[D2]], align 4
-// CHECK3-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK3-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP4]], align 4
-// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK3-NEXT: store ptr [[C1]], ptr [[TMP6]], align 4
-// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
-// CHECK3-NEXT: store ptr [[D2]], ptr [[TMP7]], align 4
+// CHECK3-NEXT: [[TMP4:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK3-NEXT: store ptr [[C1]], ptr [[TMP4]], align 4
+// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
+// CHECK3-NEXT: store ptr [[D2]], ptr [[TMP5]], align 4
// CHECK3-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// CHECK3-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP5]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 2048, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2, ptr @_omp_reduction_list_to_global_copy_func3, ptr @_omp_reduction_list_to_global_reduce_func4, ptr @_omp_reduction_global_to_list_copy_func5, ptr @_omp_reduction_global_to_list_reduce_func6)
-// CHECK3-NEXT: [[TMP9:%.*]] = icmp eq i32 [[TMP8]], 1
-// CHECK3-NEXT: br i1 [[TMP9]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK3-NEXT: [[TMP6:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 2048, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func1, ptr @_omp_reduction_inter_warp_copy_func2, ptr @_omp_reduction_list_to_global_copy_func3, ptr @_omp_reduction_list_to_global_reduce_func4, ptr @_omp_reduction_global_to_list_copy_func5, ptr @_omp_reduction_global_to_list_reduce_func6)
+// CHECK3-NEXT: [[TMP7:%.*]] = icmp eq i32 [[TMP6]], 1
+// CHECK3-NEXT: br i1 [[TMP7]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK3: .omp.reduction.then:
-// CHECK3-NEXT: [[TMP10:%.*]] = load i8, ptr [[TMP0]], align 1
-// CHECK3-NEXT: [[CONV4:%.*]] = sext i8 [[TMP10]] to i32
-// CHECK3-NEXT: [[TMP11:%.*]] = load i8, ptr [[C1]], align 1
-// CHECK3-NEXT: [[CONV5:%.*]] = sext i8 [[TMP11]] to i32
+// CHECK3-NEXT: [[TMP8:%.*]] = load i8, ptr [[TMP0]], align 1
+// CHECK3-NEXT: [[CONV4:%.*]] = sext i8 [[TMP8]] to i32
+// CHECK3-NEXT: [[TMP9:%.*]] = load i8, ptr [[C1]], align 1
+// CHECK3-NEXT: [[CONV5:%.*]] = sext i8 [[TMP9]] to i32
// CHECK3-NEXT: [[XOR6:%.*]] = xor i32 [[CONV4]], [[CONV5]]
// CHECK3-NEXT: [[CONV7:%.*]] = trunc i32 [[XOR6]] to i8
// CHECK3-NEXT: store i8 [[CONV7]], ptr [[TMP0]], align 1
-// CHECK3-NEXT: [[TMP12:%.*]] = load float, ptr [[TMP1]], align 4
-// CHECK3-NEXT: [[TMP13:%.*]] = load float, ptr [[D2]], align 4
-// CHECK3-NEXT: [[MUL8:%.*]] = fmul float [[TMP12]], [[TMP13]]
+// CHECK3-NEXT: [[TMP10:%.*]] = load float, ptr [[TMP1]], align 4
+// CHECK3-NEXT: [[TMP11:%.*]] = load float, ptr [[D2]], align 4
+// CHECK3-NEXT: [[MUL8:%.*]] = fmul float [[TMP10]], [[TMP11]]
// CHECK3-NEXT: store float [[MUL8]], ptr [[TMP1]], align 4
// CHECK3-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK3: .omp.reduction.done:
@@ -2967,13 +2951,13 @@ int bar(int n){
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP9:%.*]] = load i8, ptr [[TMP7]], align 1
// CHECK3-NEXT: store i8 [[TMP9]], ptr [[TMP8]], align 4
// CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 1
// CHECK3-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 4
// CHECK3-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP4]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [2048 x float], ptr [[D]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP13:%.*]] = load float, ptr [[TMP11]], align 4
// CHECK3-NEXT: store float [[TMP13]], ptr [[TMP12]], align 4
// CHECK3-NEXT: ret void
@@ -2993,11 +2977,11 @@ int bar(int n){
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK3-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2048 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK3-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP3]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x float], ptr [[D]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK3-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l26_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr [[TMP9]]) #[[ATTR4]]
@@ -3019,13 +3003,13 @@ int bar(int n){
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP9:%.*]] = load i8, ptr [[TMP8]], align 4
// CHECK3-NEXT: store i8 [[TMP9]], ptr [[TMP7]], align 1
// CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 1
// CHECK3-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 4
// CHECK3-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP4]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [2048 x float], ptr [[D]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP13:%.*]] = load float, ptr [[TMP12]], align 4
// CHECK3-NEXT: store float [[TMP13]], ptr [[TMP11]], align 4
// CHECK3-NEXT: ret void
@@ -3045,11 +3029,11 @@ int bar(int n){
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK3-NEXT: [[C:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2048 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i8], ptr [[C]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK3-NEXT: [[D:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_0]], ptr [[TMP3]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x float], ptr [[D]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x float], ptr [[D]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK3-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l26_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP9]], ptr [[DOTOMP_REDUCTION_RED_LIST]]) #[[ATTR4]]
@@ -3112,7 +3096,7 @@ int bar(int n){
// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK3-NEXT: store ptr [[B2]], ptr [[TMP7]], align 4
// CHECK3-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// CHECK3-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP5]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 2048, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func9, ptr @_omp_reduction_inter_warp_copy_func10, ptr @_omp_reduction_list_to_global_copy_func11, ptr @_omp_reduction_list_to_global_reduce_func12, ptr @_omp_reduction_global_to_list_copy_func13, ptr @_omp_reduction_global_to_list_reduce_func14)
+// CHECK3-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr @[[GLOB1]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 2048, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func9, ptr @_omp_reduction_inter_warp_copy_func10, ptr @_omp_reduction_list_to_global_copy_func11, ptr @_omp_reduction_list_to_global_reduce_func12, ptr @_omp_reduction_global_to_list_copy_func13, ptr @_omp_reduction_global_to_list_reduce_func14)
// CHECK3-NEXT: [[TMP9:%.*]] = icmp eq i32 [[TMP8]], 1
// CHECK3-NEXT: br i1 [[TMP9]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK3: .omp.reduction.then:
@@ -3175,34 +3159,32 @@ int bar(int n){
// CHECK3-NEXT: [[COND:%.*]] = phi i32 [ 99, [[COND_TRUE]] ], [ [[CONV3]], [[COND_FALSE]] ]
// CHECK3-NEXT: [[CONV4:%.*]] = trunc i32 [[COND]] to i16
// CHECK3-NEXT: store i16 [[CONV4]], ptr [[B2]], align 2
-// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
-// CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
-// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
-// CHECK3-NEXT: store ptr [[A1]], ptr [[TMP7]], align 4
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
-// CHECK3-NEXT: store ptr [[B2]], ptr [[TMP8]], align 4
-// CHECK3-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP6]], i32 2, i32 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func7, ptr @_omp_reduction_inter_warp_copy_func8)
-// CHECK3-NEXT: [[TMP10:%.*]] = icmp eq i32 [[TMP9]], 1
-// CHECK3-NEXT: br i1 [[TMP10]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
+// CHECK3-NEXT: store ptr [[A1]], ptr [[TMP5]], align 4
+// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
+// CHECK3-NEXT: store ptr [[B2]], ptr [[TMP6]], align 4
+// CHECK3-NEXT: [[TMP7:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func7, ptr @_omp_reduction_inter_warp_copy_func8)
+// CHECK3-NEXT: [[TMP8:%.*]] = icmp eq i32 [[TMP7]], 1
+// CHECK3-NEXT: br i1 [[TMP8]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK3: .omp.reduction.then:
-// CHECK3-NEXT: [[TMP11:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK3-NEXT: [[TMP12:%.*]] = load i32, ptr [[A1]], align 4
-// CHECK3-NEXT: [[OR5:%.*]] = or i32 [[TMP11]], [[TMP12]]
+// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK3-NEXT: [[TMP10:%.*]] = load i32, ptr [[A1]], align 4
+// CHECK3-NEXT: [[OR5:%.*]] = or i32 [[TMP9]], [[TMP10]]
// CHECK3-NEXT: store i32 [[OR5]], ptr [[TMP0]], align 4
-// CHECK3-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
-// CHECK3-NEXT: [[CONV6:%.*]] = sext i16 [[TMP13]] to i32
-// CHECK3-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
-// CHECK3-NEXT: [[CONV7:%.*]] = sext i16 [[TMP14]] to i32
+// CHECK3-NEXT: [[TMP11:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK3-NEXT: [[CONV6:%.*]] = sext i16 [[TMP11]] to i32
+// CHECK3-NEXT: [[TMP12:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK3-NEXT: [[CONV7:%.*]] = sext i16 [[TMP12]] to i32
// CHECK3-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[CONV6]], [[CONV7]]
// CHECK3-NEXT: br i1 [[CMP8]], label [[COND_TRUE9:%.*]], label [[COND_FALSE10:%.*]]
// CHECK3: cond.true9:
-// CHECK3-NEXT: [[TMP15:%.*]] = load i16, ptr [[TMP1]], align 2
+// CHECK3-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP1]], align 2
// CHECK3-NEXT: br label [[COND_END11:%.*]]
// CHECK3: cond.false10:
-// CHECK3-NEXT: [[TMP16:%.*]] = load i16, ptr [[B2]], align 2
+// CHECK3-NEXT: [[TMP14:%.*]] = load i16, ptr [[B2]], align 2
// CHECK3-NEXT: br label [[COND_END11]]
// CHECK3: cond.end11:
-// CHECK3-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP15]], [[COND_TRUE9]] ], [ [[TMP16]], [[COND_FALSE10]] ]
+// CHECK3-NEXT: [[COND12:%.*]] = phi i16 [ [[TMP13]], [[COND_TRUE9]] ], [ [[TMP14]], [[COND_FALSE10]] ]
// CHECK3-NEXT: store i16 [[COND12]], ptr [[TMP1]], align 2
// CHECK3-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK3: .omp.reduction.done:
@@ -3540,13 +3522,13 @@ int bar(int n){
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK3-NEXT: store i32 [[TMP9]], ptr [[TMP8]], align 4
// CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 1
// CHECK3-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 4
// CHECK3-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP4]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [2048 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP11]], align 2
// CHECK3-NEXT: store i16 [[TMP13]], ptr [[TMP12]], align 4
// CHECK3-NEXT: ret void
@@ -3566,11 +3548,11 @@ int bar(int n){
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2048 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK3-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP3]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK3-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l33_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr [[TMP9]]) #[[ATTR4]]
@@ -3592,13 +3574,13 @@ int bar(int n){
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP4]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP8]], align 4
// CHECK3-NEXT: store i32 [[TMP9]], ptr [[TMP7]], align 4
// CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[TMP3]], i32 0, i32 1
// CHECK3-NEXT: [[TMP11:%.*]] = load ptr, ptr [[TMP10]], align 4
// CHECK3-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP4]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [2048 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
+// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP5]]
// CHECK3-NEXT: [[TMP13:%.*]] = load i16, ptr [[TMP12]], align 4
// CHECK3-NEXT: store i16 [[TMP13]], ptr [[TMP11]], align 2
// CHECK3-NEXT: ret void
@@ -3618,11 +3600,11 @@ int bar(int n){
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2:%.*]], ptr [[TMP3]], i32 0, i32 0
-// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2048 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x i32], ptr [[A]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 4
// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 1
// CHECK3-NEXT: [[B:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY_2]], ptr [[TMP3]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2048 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x i16], ptr [[B]], i32 0, i32 [[TMP4]]
// CHECK3-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTADDR2]], align 4
// CHECK3-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l33_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP9]], ptr [[DOTOMP_REDUCTION_RED_LIST]]) #[[ATTR4]]
diff --git a/clang/test/OpenMP/reduction_implicit_map.cpp b/clang/test/OpenMP/reduction_implicit_map.cpp
index 93bee189c0e2161..262998e149ed21f 100644
--- a/clang/test/OpenMP/reduction_implicit_map.cpp
+++ b/clang/test/OpenMP/reduction_implicit_map.cpp
@@ -146,17 +146,15 @@ int main()
// CHECK-NEXT: store ptr [[TMP7]], ptr [[TMP]], align 8
// CHECK-NEXT: [[TMP8:%.*]] = load ptr, ptr [[TMP]], align 8
// CHECK-NEXT: store double 1.000000e+01, ptr [[TMP8]], align 8
-// CHECK-NEXT: [[TMP9:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK-NEXT: [[TMP10:%.*]] = load i32, ptr [[TMP9]], align 4
-// CHECK-NEXT: [[TMP11:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK-NEXT: store ptr [[E2]], ptr [[TMP11]], align 8
-// CHECK-NEXT: [[TMP12:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i32 [[TMP10]], i32 1, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
-// CHECK-NEXT: [[TMP13:%.*]] = icmp eq i32 [[TMP12]], 1
-// CHECK-NEXT: br i1 [[TMP13]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK-NEXT: store ptr [[E2]], ptr [[TMP9]], align 8
+// CHECK-NEXT: [[TMP10:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr @[[GLOB1]], i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
+// CHECK-NEXT: [[TMP11:%.*]] = icmp eq i32 [[TMP10]], 1
+// CHECK-NEXT: br i1 [[TMP11]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// CHECK: .omp.reduction.then:
-// CHECK-NEXT: [[TMP14:%.*]] = load double, ptr [[ARRAYIDX]], align 8
-// CHECK-NEXT: [[TMP15:%.*]] = load double, ptr [[E2]], align 8
-// CHECK-NEXT: [[ADD:%.*]] = fadd double [[TMP14]], [[TMP15]]
+// CHECK-NEXT: [[TMP12:%.*]] = load double, ptr [[ARRAYIDX]], align 8
+// CHECK-NEXT: [[TMP13:%.*]] = load double, ptr [[E2]], align 8
+// CHECK-NEXT: [[ADD:%.*]] = fadd double [[TMP12]], [[TMP13]]
// CHECK-NEXT: store double [[ADD]], ptr [[ARRAYIDX]], align 8
// CHECK-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// CHECK: .omp.reduction.done:
diff --git a/clang/test/OpenMP/target_teams_generic_loop_codegen.cpp b/clang/test/OpenMP/target_teams_generic_loop_codegen.cpp
index bc5fd39dde8d429..87d55c870ae4f5e 100644
--- a/clang/test/OpenMP/target_teams_generic_loop_codegen.cpp
+++ b/clang/test/OpenMP/target_teams_generic_loop_codegen.cpp
@@ -1322,28 +1322,26 @@ int foo() {
// IR-GPU-NEXT: store i32 [[TMP36]], ptr [[J_ADDR_ASCAST]], align 4
// IR-GPU-NEXT: br label [[DOTOMP_LASTPRIVATE_DONE]]
// IR-GPU: .omp.lastprivate.done:
-// IR-GPU-NEXT: [[TMP37:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
-// IR-GPU-NEXT: [[TMP38:%.*]] = load i32, ptr [[TMP37]], align 4
-// IR-GPU-NEXT: [[TMP39:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], i64 0, i64 0
-// IR-GPU-NEXT: store ptr [[SUM1_ASCAST]], ptr [[TMP39]], align 8
+// IR-GPU-NEXT: [[TMP37:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], i64 0, i64 0
+// IR-GPU-NEXT: store ptr [[SUM1_ASCAST]], ptr [[TMP37]], align 8
// IR-GPU-NEXT: %"_openmp_teams_reductions_buffer_$_$ptr" = call ptr @__kmpc_reduction_get_fixed_buffer()
-// IR-GPU-NEXT: [[TMP40:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr addrspacecast (ptr addrspace(1) @[[GLOB1]] to ptr), i32 [[TMP38]], ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], ptr @_omp_reduction_shuffle_and_reduce_func.1, ptr @_omp_reduction_inter_warp_copy_func.2, ptr @_omp_reduction_list_to_global_copy_func, ptr @_omp_reduction_list_to_global_reduce_func, ptr @_omp_reduction_global_to_list_copy_func, ptr @_omp_reduction_global_to_list_reduce_func)
-// IR-GPU-NEXT: [[TMP41:%.*]] = icmp eq i32 [[TMP40]], 1
-// IR-GPU-NEXT: br i1 [[TMP41]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// IR-GPU-NEXT: [[TMP38:%.*]] = call i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr addrspacecast (ptr addrspace(1) @[[GLOB1]] to ptr), ptr %"_openmp_teams_reductions_buffer_$_$ptr", i32 1024, i64 400, ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], ptr @_omp_reduction_shuffle_and_reduce_func.1, ptr @_omp_reduction_inter_warp_copy_func.2, ptr @_omp_reduction_list_to_global_copy_func, ptr @_omp_reduction_list_to_global_reduce_func, ptr @_omp_reduction_global_to_list_copy_func, ptr @_omp_reduction_global_to_list_reduce_func)
+// IR-GPU-NEXT: [[TMP39:%.*]] = icmp eq i32 [[TMP38]], 1
+// IR-GPU-NEXT: br i1 [[TMP39]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// IR-GPU: .omp.reduction.then:
-// IR-GPU-NEXT: [[TMP42:%.*]] = getelementptr i32, ptr [[TMP0]], i64 100
-// IR-GPU-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[TMP0]], [[TMP42]]
+// IR-GPU-NEXT: [[TMP40:%.*]] = getelementptr i32, ptr [[TMP0]], i64 100
+// IR-GPU-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[TMP0]], [[TMP40]]
// IR-GPU-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE17:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// IR-GPU: omp.arraycpy.body:
// IR-GPU-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[SUM1_ASCAST]], [[DOTOMP_REDUCTION_THEN]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// IR-GPU-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST13:%.*]] = phi ptr [ [[TMP0]], [[DOTOMP_REDUCTION_THEN]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT15:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// IR-GPU-NEXT: [[TMP43:%.*]] = load i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST13]], align 4
-// IR-GPU-NEXT: [[TMP44:%.*]] = load i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 4
-// IR-GPU-NEXT: [[ADD14:%.*]] = add nsw i32 [[TMP43]], [[TMP44]]
+// IR-GPU-NEXT: [[TMP41:%.*]] = load i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST13]], align 4
+// IR-GPU-NEXT: [[TMP42:%.*]] = load i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 4
+// IR-GPU-NEXT: [[ADD14:%.*]] = add nsw i32 [[TMP41]], [[TMP42]]
// IR-GPU-NEXT: store i32 [[ADD14]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST13]], align 4
// IR-GPU-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT15]] = getelementptr i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST13]], i32 1
// IR-GPU-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// IR-GPU-NEXT: [[OMP_ARRAYCPY_DONE16:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT15]], [[TMP42]]
+// IR-GPU-NEXT: [[OMP_ARRAYCPY_DONE16:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT15]], [[TMP40]]
// IR-GPU-NEXT: br i1 [[OMP_ARRAYCPY_DONE16]], label [[OMP_ARRAYCPY_DONE17]], label [[OMP_ARRAYCPY_BODY]]
// IR-GPU: omp.arraycpy.done17:
// IR-GPU-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
@@ -1469,38 +1467,36 @@ int foo() {
// IR-GPU-NEXT: [[TMP18:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
// IR-GPU-NEXT: [[TMP19:%.*]] = load i32, ptr [[TMP18]], align 4
// IR-GPU-NEXT: call void @__kmpc_for_static_fini(ptr addrspacecast (ptr addrspace(1) @[[GLOB3]] to ptr), i32 [[TMP19]])
-// IR-GPU-NEXT: [[TMP20:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
-// IR-GPU-NEXT: [[TMP21:%.*]] = load i32, ptr [[TMP20]], align 4
-// IR-GPU-NEXT: [[TMP22:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], i64 0, i64 0
-// IR-GPU-NEXT: store ptr [[SUM4_ASCAST]], ptr [[TMP22]], align 8
-// IR-GPU-NEXT: [[TMP23:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr addrspacecast (ptr addrspace(1) @[[GLOB1]] to ptr), i32 [[TMP21]], i32 1, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
-// IR-GPU-NEXT: [[TMP24:%.*]] = icmp eq i32 [[TMP23]], 1
-// IR-GPU-NEXT: br i1 [[TMP24]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
+// IR-GPU-NEXT: [[TMP20:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], i64 0, i64 0
+// IR-GPU-NEXT: store ptr [[SUM4_ASCAST]], ptr [[TMP20]], align 8
+// IR-GPU-NEXT: [[TMP21:%.*]] = call i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr addrspacecast (ptr addrspace(1) @[[GLOB1]] to ptr), i64 400, ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], ptr @_omp_reduction_shuffle_and_reduce_func, ptr @_omp_reduction_inter_warp_copy_func)
+// IR-GPU-NEXT: [[TMP22:%.*]] = icmp eq i32 [[TMP21]], 1
+// IR-GPU-NEXT: br i1 [[TMP22]], label [[DOTOMP_REDUCTION_THEN:%.*]], label [[DOTOMP_REDUCTION_DONE:%.*]]
// IR-GPU: .omp.reduction.then:
-// IR-GPU-NEXT: [[TMP25:%.*]] = getelementptr i32, ptr [[TMP0]], i64 100
-// IR-GPU-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[TMP0]], [[TMP25]]
+// IR-GPU-NEXT: [[TMP23:%.*]] = getelementptr i32, ptr [[TMP0]], i64 100
+// IR-GPU-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[TMP0]], [[TMP23]]
// IR-GPU-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE19:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// IR-GPU: omp.arraycpy.body:
// IR-GPU-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[SUM4_ASCAST]], [[DOTOMP_REDUCTION_THEN]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// IR-GPU-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST15:%.*]] = phi ptr [ [[TMP0]], [[DOTOMP_REDUCTION_THEN]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT17:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// IR-GPU-NEXT: [[TMP26:%.*]] = load i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], align 4
-// IR-GPU-NEXT: [[TMP27:%.*]] = load i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 4
-// IR-GPU-NEXT: [[ADD16:%.*]] = add nsw i32 [[TMP26]], [[TMP27]]
+// IR-GPU-NEXT: [[TMP24:%.*]] = load i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], align 4
+// IR-GPU-NEXT: [[TMP25:%.*]] = load i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 4
+// IR-GPU-NEXT: [[ADD16:%.*]] = add nsw i32 [[TMP24]], [[TMP25]]
// IR-GPU-NEXT: store i32 [[ADD16]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], align 4
// IR-GPU-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT17]] = getelementptr i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], i32 1
// IR-GPU-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// IR-GPU-NEXT: [[OMP_ARRAYCPY_DONE18:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT17]], [[TMP25]]
+// IR-GPU-NEXT: [[OMP_ARRAYCPY_DONE18:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT17]], [[TMP23]]
// IR-GPU-NEXT: br i1 [[OMP_ARRAYCPY_DONE18]], label [[OMP_ARRAYCPY_DONE19]], label [[OMP_ARRAYCPY_BODY]]
// IR-GPU: omp.arraycpy.done19:
// IR-GPU-NEXT: br label [[DOTOMP_REDUCTION_DONE]]
// IR-GPU: .omp.reduction.done:
-// IR-GPU-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTOMP_IS_LAST_ASCAST]], align 4
-// IR-GPU-NEXT: [[TMP29:%.*]] = icmp ne i32 [[TMP28]], 0
-// IR-GPU-NEXT: br i1 [[TMP29]], label [[DOTOMP_LASTPRIVATE_THEN:%.*]], label [[DOTOMP_LASTPRIVATE_DONE:%.*]]
+// IR-GPU-NEXT: [[TMP26:%.*]] = load i32, ptr [[DOTOMP_IS_LAST_ASCAST]], align 4
+// IR-GPU-NEXT: [[TMP27:%.*]] = icmp ne i32 [[TMP26]], 0
+// IR-GPU-NEXT: br i1 [[TMP27]], label [[DOTOMP_LASTPRIVATE_THEN:%.*]], label [[DOTOMP_LASTPRIVATE_DONE:%.*]]
// IR-GPU: .omp.lastprivate.then:
// IR-GPU-NEXT: store i32 10, ptr [[J3_ASCAST]], align 4
-// IR-GPU-NEXT: [[TMP30:%.*]] = load i32, ptr [[J3_ASCAST]], align 4
-// IR-GPU-NEXT: store i32 [[TMP30]], ptr [[J_ADDR_ASCAST]], align 4
+// IR-GPU-NEXT: [[TMP28:%.*]] = load i32, ptr [[J3_ASCAST]], align 4
+// IR-GPU-NEXT: store i32 [[TMP28]], ptr [[J_ADDR_ASCAST]], align 4
// IR-GPU-NEXT: br label [[DOTOMP_LASTPRIVATE_DONE]]
// IR-GPU: .omp.lastprivate.done:
// IR-GPU-NEXT: ret void
@@ -1814,7 +1810,7 @@ int foo() {
// IR-GPU-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i64 0, i64 0
// IR-GPU-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// IR-GPU-NEXT: [[SUM:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP4]], i32 0, i32 0
-// IR-GPU-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x [10 x [10 x i32]]], ptr [[SUM]], i32 0, i32 [[TMP5]]
+// IR-GPU-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x [10 x [10 x i32]]], ptr [[SUM]], i32 0, i32 [[TMP5]]
// IR-GPU-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[TMP8]], ptr align 4 [[TMP7]], i64 400, i1 false)
// IR-GPU-NEXT: ret void
//
@@ -1837,7 +1833,7 @@ int foo() {
// IR-GPU-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1_ASCAST]], align 4
// IR-GPU-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], i64 0, i64 0
// IR-GPU-NEXT: [[SUM:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP3]], i32 0, i32 0
-// IR-GPU-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x [10 x [10 x i32]]], ptr [[SUM]], i32 0, i32 [[TMP4]]
+// IR-GPU-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x [10 x [10 x i32]]], ptr [[SUM]], i32 0, i32 [[TMP4]]
// IR-GPU-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 8
// IR-GPU-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTADDR2_ASCAST]], align 8
// IR-GPU-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3foov_l22_omp_outlined_omp$reduction$reduction_func"(ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], ptr [[TMP7]]) #[[ATTR2]]
@@ -1862,7 +1858,7 @@ int foo() {
// IR-GPU-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i64 0, i64 0
// IR-GPU-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// IR-GPU-NEXT: [[SUM:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP4]], i32 0, i32 0
-// IR-GPU-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1024 x [10 x [10 x i32]]], ptr [[SUM]], i32 0, i32 [[TMP5]]
+// IR-GPU-NEXT: [[TMP8:%.*]] = getelementptr inbounds [1 x [10 x [10 x i32]]], ptr [[SUM]], i32 0, i32 [[TMP5]]
// IR-GPU-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[TMP7]], ptr align 4 [[TMP8]], i64 400, i1 false)
// IR-GPU-NEXT: ret void
//
@@ -1885,7 +1881,7 @@ int foo() {
// IR-GPU-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTADDR1_ASCAST]], align 4
// IR-GPU-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]], i64 0, i64 0
// IR-GPU-NEXT: [[SUM:%.*]] = getelementptr inbounds [[STRUCT__GLOBALIZED_LOCALS_TY:%.*]], ptr [[TMP3]], i32 0, i32 0
-// IR-GPU-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1024 x [10 x [10 x i32]]], ptr [[SUM]], i32 0, i32 [[TMP4]]
+// IR-GPU-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x [10 x [10 x i32]]], ptr [[SUM]], i32 0, i32 [[TMP4]]
// IR-GPU-NEXT: store ptr [[TMP6]], ptr [[TMP5]], align 8
// IR-GPU-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTADDR2_ASCAST]], align 8
// IR-GPU-NEXT: call void @"{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3foov_l22_omp_outlined_omp$reduction$reduction_func"(ptr [[TMP7]], ptr [[DOTOMP_REDUCTION_RED_LIST_ASCAST]]) #[[ATTR2]]
diff --git a/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h b/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
index dc9a6497140ddae..7fdd7bf868b8bbe 100644
--- a/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
+++ b/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
@@ -2039,10 +2039,13 @@ class OpenMPIRBuilder {
/// Create a runtime call for kmpc_target_deinit
///
/// \param Loc The insert and source location description.
- /// \param TeamsReductionBufferSize The size to be allocated for the teams
- /// reduction buffer.
+ /// \param TeamsReductionDataSize The maximal size of all the reduction data
+ /// for teams reduction.
+ /// \param TeamsReductionBufferLength The number of elements (each of up to
+ /// \p TeamsReductionDataSize size), in the teams reduction buffer.
void createTargetDeinit(const LocationDescription &Loc,
- int32_t TeamsReductionBufferSize = 0);
+ int32_t TeamsReductionDataSize = 0,
+ int32_t TeamsReductionBufferLength = 1024);
///}
diff --git a/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def b/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
index 5215a5a97a0cbdc..6a719d231142230 100644
--- a/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
+++ b/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
@@ -95,7 +95,7 @@ __OMP_STRUCT_TYPE(AsyncInfo, __tgt_async_info, false, Int8Ptr)
__OMP_STRUCT_TYPE(DependInfo, kmp_dep_info, false, SizeTy, SizeTy, Int8)
__OMP_STRUCT_TYPE(Task, kmp_task_ompbuilder_t, false, VoidPtr, VoidPtr, Int32, VoidPtr, VoidPtr)
__OMP_STRUCT_TYPE(ConfigurationEnvironment, ConfigurationEnvironmentTy, false,
- Int8, Int8, Int8, Int32, Int32, Int32, Int32, Int32)
+ Int8, Int8, Int8, Int32, Int32, Int32, Int32, Int32, Int32)
__OMP_STRUCT_TYPE(DynamicEnvironment, DynamicEnvironmentTy, false, Int16)
__OMP_STRUCT_TYPE(KernelEnvironment, KernelEnvironmentTy, false,
ConfigurationEnvironment, IdentPtr, DynamicEnvironmentPtr)
@@ -469,10 +469,10 @@ __OMP_RTL(__kmpc_kernel_end_parallel, false, Void, )
__OMP_RTL(__kmpc_serialized_parallel, false, Void, IdentPtr, Int32)
__OMP_RTL(__kmpc_end_serialized_parallel, false, Void, IdentPtr, Int32)
__OMP_RTL(__kmpc_shuffle_int32, false, Int32, Int32, Int16, Int16)
-__OMP_RTL(__kmpc_nvptx_parallel_reduce_nowait_v2, false, Int32, IdentPtr, Int32,
- Int32, SizeTy, VoidPtr, ShuffleReducePtr, InterWarpCopyPtr)
-__OMP_RTL(__kmpc_nvptx_teams_reduce_nowait_v2, false, Int32, IdentPtr, Int32,
- VoidPtr, Int32, VoidPtr, ShuffleReducePtr, InterWarpCopyPtr,
+__OMP_RTL(__kmpc_nvptx_parallel_reduce_nowait_v2, false, Int32, IdentPtr,
+ Int64, VoidPtr, ShuffleReducePtr, InterWarpCopyPtr)
+__OMP_RTL(__kmpc_nvptx_teams_reduce_nowait_v2, false, Int32, IdentPtr,
+ VoidPtr, Int32, Int64, VoidPtr, ShuffleReducePtr, InterWarpCopyPtr,
GlobalListPtr, GlobalListPtr, GlobalListPtr, GlobalListPtr)
__OMP_RTL(__kmpc_reduction_get_fixed_buffer, false, VoidPtr, )
@@ -1040,9 +1040,9 @@ __OMP_RTL_ATTRS(__kmpc_end_serialized_parallel, InaccessibleArgOnlyAttrs,
__OMP_RTL_ATTRS(__kmpc_shuffle_int32, AttributeSet(), SExt,
ParamAttrs(SExt, SExt, SExt))
__OMP_RTL_ATTRS(__kmpc_nvptx_parallel_reduce_nowait_v2, AttributeSet(), SExt,
- ParamAttrs(AttributeSet(), SExt, SExt, SizeTyExt))
+ ParamAttrs())
__OMP_RTL_ATTRS(__kmpc_nvptx_teams_reduce_nowait_v2, AttributeSet(), SExt,
- ParamAttrs(AttributeSet(), SExt, AttributeSet(), ZExt))
+ ParamAttrs(AttributeSet(), AttributeSet(), ZExt))
__OMP_RTL_ATTRS(__kmpc_reduction_get_fixed_buffer, GetterAttrs, AttributeSet(), ParamAttrs())
__OMP_RTL_ATTRS(__kmpc_shuffle_int64, AttributeSet(), AttributeSet(),
diff --git a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
index 6b7ab48a93130d6..9f0e0ad57c4669b 100644
--- a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
+++ b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
@@ -4251,7 +4251,8 @@ OpenMPIRBuilder::createTargetInit(const LocationDescription &Loc, bool IsSPMD,
Constant *MaxThreads = ConstantInt::getSigned(Int32, MaxThreadsVal);
Constant *MinTeams = ConstantInt::getSigned(Int32, MinTeamsVal);
Constant *MaxTeams = ConstantInt::getSigned(Int32, MaxTeamsVal);
- Constant *ReductionBufferSize = ConstantInt::getSigned(Int32, 0);
+ Constant *ReductionDataSize = ConstantInt::getSigned(Int32, 0);
+ Constant *ReductionBufferLength = ConstantInt::getSigned(Int32, 0);
// We need to strip the debug prefix to get the correct kernel name.
StringRef KernelName = Kernel->getName();
@@ -4288,7 +4289,8 @@ OpenMPIRBuilder::createTargetInit(const LocationDescription &Loc, bool IsSPMD,
MaxThreads,
MinTeams,
MaxTeams,
- ReductionBufferSize,
+ ReductionDataSize,
+ ReductionBufferLength,
});
Constant *KernelEnvironmentInitializer = ConstantStruct::get(
KernelEnvironment, {
@@ -4345,7 +4347,8 @@ OpenMPIRBuilder::createTargetInit(const LocationDescription &Loc, bool IsSPMD,
}
void OpenMPIRBuilder::createTargetDeinit(const LocationDescription &Loc,
- int32_t TeamsReductionBufferSize) {
+ int32_t TeamsReductionDataSize,
+ int32_t TeamsReductionBufferLength) {
if (!updateToLocation(Loc))
return;
@@ -4354,7 +4357,7 @@ void OpenMPIRBuilder::createTargetDeinit(const LocationDescription &Loc,
Builder.CreateCall(Fn, {});
- if (!TeamsReductionBufferSize)
+ if (!TeamsReductionBufferLength || !TeamsReductionDataSize)
return;
Function *Kernel = Builder.GetInsertBlock()->getParent();
@@ -4369,7 +4372,10 @@ void OpenMPIRBuilder::createTargetDeinit(const LocationDescription &Loc,
auto *KernelEnvironmentInitializer = KernelEnvironmentGV->getInitializer();
auto *NewInitializer = ConstantFoldInsertValueInstruction(
KernelEnvironmentInitializer,
- ConstantInt::get(Int32, TeamsReductionBufferSize), {0, 7});
+ ConstantInt::get(Int32, TeamsReductionDataSize), {0, 7});
+ NewInitializer = ConstantFoldInsertValueInstruction(
+ NewInitializer, ConstantInt::get(Int32, TeamsReductionBufferLength),
+ {0, 8});
KernelEnvironmentGV->setInitializer(NewInitializer);
}
diff --git a/llvm/test/Transforms/OpenMP/add_attributes.ll b/llvm/test/Transforms/OpenMP/add_attributes.ll
index 18271b7646bc97d..47ff5cad4e7eb4c 100644
--- a/llvm/test/Transforms/OpenMP/add_attributes.ll
+++ b/llvm/test/Transforms/OpenMP/add_attributes.ll
@@ -722,9 +722,9 @@ declare void @__kmpc_kernel_prepare_parallel(ptr);
declare i32 @__kmpc_masked(ptr, i32, i32);
-declare i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr, i32, i32, i64, ptr, ptr, ptr);
+declare i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr, i64, ptr, ptr, ptr);
-declare i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr, i32, ptr, i32, ptr, ptr, ptr, ptr, ptr, ptr, ptr);
+declare i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr, ptr, i32, i64, ptr, ptr, ptr, ptr, ptr, ptr, ptr);
declare i32 @__kmpc_omp_reg_task_with_affinity(ptr, i32, ptr, i32, ptr);
@@ -1366,10 +1366,10 @@ declare i32 @__tgt_target_kernel_nowait(ptr, i64, i32, i32, ptr, ptr, i32, ptr,
; CHECK: declare i32 @__kmpc_masked(ptr, i32, i32)
; CHECK-NOT: Function Attrs
-; CHECK: declare i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr, i32, i32, i64, ptr, ptr, ptr)
+; CHECK: declare i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr, i64, ptr, ptr, ptr)
; CHECK-NOT: Function Attrs
-; CHECK: declare i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr, i32, ptr, i32, ptr, ptr, ptr, ptr, ptr, ptr, ptr)
+; CHECK: declare i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr, ptr, i32, i64, ptr, ptr, ptr, ptr, ptr, ptr, ptr)
; CHECK: ; Function Attrs: nounwind
; CHECK: declare i32 @__kmpc_omp_reg_task_with_affinity(ptr, i32, ptr, i32, ptr)
@@ -2011,10 +2011,10 @@ declare i32 @__tgt_target_kernel_nowait(ptr, i64, i32, i32, ptr, ptr, i32, ptr,
; OPTIMISTIC: declare i32 @__kmpc_masked(ptr nocapture nofree readonly, i32, i32)
; OPTIMISTIC-NOT: Function Attrs
-; OPTIMISTIC: declare i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr, i32, i32, i64, ptr, ptr, ptr)
+; OPTIMISTIC: declare i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr, i64, ptr, ptr, ptr)
; OPTIMISTIC-NOT: Function Attrs
-; OPTIMISTIC: declare i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr, i32, ptr, i32, ptr, ptr, ptr, ptr, ptr, ptr, ptr)
+; OPTIMISTIC: declare i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr, ptr, i32, i64, ptr, ptr, ptr, ptr, ptr, ptr, ptr)
; OPTIMISTIC: nofree nosync nounwind willreturn
; OPTIMISTIC: declare i32 @__kmpc_omp_reg_task_with_affinity(ptr nocapture nofree readonly, i32, ptr nocapture nofree readonly, i32, ptr nocapture nofree readonly)
@@ -2669,10 +2669,10 @@ declare i32 @__tgt_target_kernel_nowait(ptr, i64, i32, i32, ptr, ptr, i32, ptr,
; EXT: declare signext i32 @__kmpc_masked(ptr, i32 signext, i32 signext)
; EXT-NOT: Function Attrs
-; EXT: declare signext i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr, i32 signext, i32 signext, i64, ptr, ptr, ptr)
+; EXT: declare signext i32 @__kmpc_nvptx_parallel_reduce_nowait_v2(ptr, i64, ptr, ptr, ptr)
; EXT-NOT: Function Attrs
-; EXT: declare signext i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr, i32 signext, ptr, i32 zeroext, ptr, ptr, ptr, ptr, ptr, ptr, ptr)
+; EXT: declare signext i32 @__kmpc_nvptx_teams_reduce_nowait_v2(ptr, ptr, i32 zeroext, i64, ptr, ptr, ptr, ptr, ptr, ptr, ptr)
; EXT: ; Function Attrs: nounwind
; EXT: declare signext i32 @__kmpc_omp_reg_task_with_affinity(ptr, i32 signext, ptr, i32 signext, ptr)
diff --git a/llvm/test/Transforms/OpenMP/always_inline_device.ll b/llvm/test/Transforms/OpenMP/always_inline_device.ll
index 2e4b8d76795a0f3..4641aff1ea21d95 100644
--- a/llvm/test/Transforms/OpenMP/always_inline_device.ll
+++ b/llvm/test/Transforms/OpenMP/always_inline_device.ll
@@ -3,19 +3,19 @@
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@G = external global i8
- at kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
; Function Attrs: convergent norecurse nounwind
;.
; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i8
-; CHECK: @[[KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
define weak void @__omp_offloading_fd02_c0934fc2_foo_l4(ptr %dyn) #0 {
; CHECK: Function Attrs: norecurse nounwind
diff --git a/llvm/test/Transforms/OpenMP/custom_state_machines.ll b/llvm/test/Transforms/OpenMP/custom_state_machines.ll
index 47a580e13b9181f..1a6624b46f10dc7 100644
--- a/llvm/test/Transforms/OpenMP/custom_state_machines.ll
+++ b/llvm/test/Transforms/OpenMP/custom_state_machines.ll
@@ -121,7 +121,7 @@
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@@ -129,14 +129,14 @@
@G = external global i32, align 4
@3 = private unnamed_addr constant %struct.ident_t { i32 0, i32 322, i32 2, i32 0, ptr @0 }, align 8
- at __omp_offloading_14_a36502b_no_state_machine_needed_l14_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_l22_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_interprocedural_l39_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_with_fallback_l55_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_no_openmp_attr_l66_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_pure_l77_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_interprocedural_nested_recursive_l92_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_no_state_machine_weak_callee_l112_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_no_state_machine_needed_l14_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_l22_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_interprocedural_l39_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_with_fallback_l55_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_no_openmp_attr_l66_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_pure_l77_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_interprocedural_nested_recursive_l92_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_no_state_machine_weak_callee_l112_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
define weak void @__omp_offloading_14_a36502b_no_state_machine_needed_l14(ptr %dyn) #0 {
entry:
@@ -840,14 +840,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; AMDGPU: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; AMDGPU: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; AMDGPU: @[[__OMP_OUTLINED__2_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
; AMDGPU: @[[__OMP_OUTLINED__3_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
; AMDGPU: @[[__OMP_OUTLINED__5_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
@@ -863,14 +863,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; NVPTX: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; NVPTX: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; NVPTX: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; NVPTX: @[[__OMP_OUTLINED__2_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
; NVPTX: @[[__OMP_OUTLINED__3_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
; NVPTX: @[[__OMP_OUTLINED__5_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
@@ -886,28 +886,28 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU-DISABLED: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; AMDGPU-DISABLED: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; AMDGPU-DISABLED: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; NVPTX-DISABLED: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; NVPTX-DISABLED: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; NVPTX-DISABLED: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; NVPTX-DISABLED: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; NVPTX-DISABLED: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; AMDGPU: Function Attrs: convergent noinline norecurse nounwind
; AMDGPU-LABEL: define {{[^@]+}}@__omp_offloading_14_a36502b_no_state_machine_needed_l14
diff --git a/llvm/test/Transforms/OpenMP/custom_state_machines_pre_lto.ll b/llvm/test/Transforms/OpenMP/custom_state_machines_pre_lto.ll
index c60d07ada2fd09e..fe134ce350dc83f 100644
--- a/llvm/test/Transforms/OpenMP/custom_state_machines_pre_lto.ll
+++ b/llvm/test/Transforms/OpenMP/custom_state_machines_pre_lto.ll
@@ -123,21 +123,21 @@
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@2 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 2, i32 0, ptr @0 }, align 8
@G = external global i32, align 4
@3 = private unnamed_addr constant %struct.ident_t { i32 0, i32 322, i32 2, i32 0, ptr @0 }, align 8
- at __omp_offloading_14_a36502b_no_state_machine_needed_l14_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_l22_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_interprocedural_l39_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_with_fallback_l55_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_no_openmp_attr_l66_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_pure_l77_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_simple_state_machine_interprocedural_nested_recursive_l92_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_14_a36502b_no_state_machine_weak_callee_l112_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_no_state_machine_needed_l14_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_l22_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_interprocedural_l39_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_with_fallback_l55_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_no_openmp_attr_l66_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_pure_l77_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_simple_state_machine_interprocedural_nested_recursive_l92_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_14_a36502b_no_state_machine_weak_callee_l112_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
define weak void @__omp_offloading_14_a36502b_no_state_machine_needed_l14(ptr %dyn) #0 {
entry:
@@ -839,84 +839,84 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU1: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; AMDGPU1: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; AMDGPU1: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU1: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; NVPTX1: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; NVPTX1: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; NVPTX1: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; NVPTX1: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; NVPTX1: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX1: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; AMDGPU2: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; AMDGPU2: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; AMDGPU2: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; AMDGPU2: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; AMDGPU2: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU2: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; AMDGPU3: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; AMDGPU3: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; AMDGPU3: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; AMDGPU3: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; AMDGPU3: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU3: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; NVPTX2: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; NVPTX2: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; NVPTX2: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; NVPTX2: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; NVPTX2: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX2: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; NVPTX3: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; NVPTX3: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; NVPTX3: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; NVPTX3: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
; NVPTX3: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 322, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
-; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_NEEDED_L14_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_L22_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_L39_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_WITH_FALLBACK_L55_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_NO_OPENMP_ATTR_L66_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_PURE_L77_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_SIMPLE_STATE_MACHINE_INTERPROCEDURAL_NESTED_RECURSIVE_L92_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX3: @[[__OMP_OFFLOADING_14_A36502B_NO_STATE_MACHINE_WEAK_CALLEE_L112_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; AMDGPU1: Function Attrs: convergent noinline norecurse nounwind
; AMDGPU1-LABEL: define {{[^@]+}}@__omp_offloading_14_a36502b_no_state_machine_needed_l14
diff --git a/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll b/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll
index 432c853fcaa6740..f8c4e6b113c9b42 100644
--- a/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll
+++ b/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll
@@ -37,7 +37,7 @@ target triple = "nvptx64"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [113 x i8] c";llvm/test/Transforms/OpenMP/custom_state_machines_remarks.c;__omp_offloading_2a_d80d3d_test_fallback_l11;11;1;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@@ -54,8 +54,8 @@ target triple = "nvptx64"
@12 = private unnamed_addr constant [73 x i8] c";llvm/test/Transforms/OpenMP/custom_state_machines_remarks.c;known;4;1;;\00", align 1
@13 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 2, i32 0, ptr @12 }, align 8
@G = external global i32
- at __omp_offloading_2a_d80d3d_test_fallback_l11_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
- at __omp_offloading_2a_d80d3d_test_no_fallback_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at __omp_offloading_2a_d80d3d_test_fallback_l11_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at __omp_offloading_2a_d80d3d_test_no_fallback_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
; Function Attrs: convergent norecurse nounwind
diff --git a/llvm/test/Transforms/OpenMP/deduplication_target.ll b/llvm/test/Transforms/OpenMP/deduplication_target.ll
index a5d54414f0086ab..6b0563365c64808 100644
--- a/llvm/test/Transforms/OpenMP/deduplication_target.ll
+++ b/llvm/test/Transforms/OpenMP/deduplication_target.ll
@@ -6,12 +6,12 @@ target triple = "nvptx64"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@2 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 1, i32 0, ptr @0 }, align 8
- at __omp_offloading_50_a3e09bf8_foo_l2_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at __omp_offloading_50_a3e09bf8_foo_l2_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
declare void @use(i32)
diff --git a/llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll b/llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll
index 7de8fa65fcc05b4..078ec22e611077b 100644
--- a/llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll
+++ b/llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll
@@ -3,19 +3,19 @@
target triple = "nvptx64"
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@G = external global i32
- at kernel0_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
- at kernel1_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
- at kernel2_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at kernel0_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at kernel1_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at kernel2_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
;.
; CHECK: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32
-; CHECK: @[[KERNEL0_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 666, i32 0, i32 777 }, ptr null, ptr null }
-; CHECK: @[[KERNEL1_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 666, i32 0, i32 777 }, ptr null, ptr null }
-; CHECK: @[[KERNEL2_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 666, i32 0, i32 777 }, ptr null, ptr null }
+; CHECK: @[[KERNEL0_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 666, i32 0, i32 777, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[KERNEL1_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 666, i32 0, i32 777, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[KERNEL2_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 666, i32 0, i32 777, i32 0, i32 0 }, ptr null, ptr null }
; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 22, ptr @[[GLOB0]] }, align 8
;.
diff --git a/llvm/test/Transforms/OpenMP/global_constructor.ll b/llvm/test/Transforms/OpenMP/global_constructor.ll
index 7ac04f026812058..804b910dcd3088b 100644
--- a/llvm/test/Transforms/OpenMP/global_constructor.ll
+++ b/llvm/test/Transforms/OpenMP/global_constructor.ll
@@ -3,12 +3,12 @@
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@_ZL6Device = internal global double 0.000000e+00, align 8
- at __omp_offloading_fd02_85283c04_main_l11_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_fd02_85283c04_main_l11_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
define weak void @__omp_offloading_fd02_85283c04_main_l11(ptr %dyn, ptr nonnull align 8 dereferenceable(8) %X) local_unnamed_addr "kernel" {
entry:
diff --git a/llvm/test/Transforms/OpenMP/globalization_remarks.ll b/llvm/test/Transforms/OpenMP/globalization_remarks.ll
index 9a44343b5f52463..878ac9010a7dc3f 100644
--- a/llvm/test/Transforms/OpenMP/globalization_remarks.ll
+++ b/llvm/test/Transforms/OpenMP/globalization_remarks.ll
@@ -8,10 +8,10 @@ target triple = "nvptx64"
; CHECK: remark: globalization_remarks.c:5:7: Found thread data sharing on the GPU. Expect degraded performance due to data globalization.
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@S = external local_unnamed_addr global ptr
- at foo_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at foo_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
define void @foo() "kernel" {
entry:
diff --git a/llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll b/llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll
index 9363b29d00e245c..34c7113eda1685a 100644
--- a/llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll
+++ b/llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll
@@ -37,12 +37,12 @@
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@2 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 2, i32 0, ptr @0 }, align 8
- at __omp_offloading_10301_87b2c_foo_l7_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_10301_87b2c_foo_l7_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
define weak void @__omp_offloading_10301_87b2c_foo_l7() "kernel" {
entry:
diff --git a/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll b/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll
index f0feec51d3c2dfa..acaacf8af0ac9a2 100644
--- a/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll
+++ b/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll
@@ -3,20 +3,20 @@
target triple = "nvptx64"
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@G = external global i8
- at is_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
- at will_be_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
- at none_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
- at will_not_be_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at is_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at will_be_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at none_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at will_not_be_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
;.
; CHECK: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i8
-; CHECK: @[[IS_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
-; CHECK: @[[WILL_BE_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
-; CHECK: @[[NONE_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
-; CHECK: @[[WILL_NOT_BE_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[IS_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[WILL_BE_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[NONE_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[WILL_NOT_BE_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
;.
define weak void @is_spmd() "kernel" {
; CHECK-LABEL: define {{[^@]+}}@is_spmd
diff --git a/llvm/test/Transforms/OpenMP/nested_parallelism.ll b/llvm/test/Transforms/OpenMP/nested_parallelism.ll
index fafd05f4efb9b0e..adc9c3fd311d453 100644
--- a/llvm/test/Transforms/OpenMP/nested_parallelism.ll
+++ b/llvm/test/Transforms/OpenMP/nested_parallelism.ll
@@ -24,15 +24,15 @@ target triple = "nvptx64"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 22, ptr @0 }, align 8
@i_shared = internal addrspace(3) global [4 x i8] undef, align 16
@i.i_shared = internal addrspace(3) global [4 x i8] undef, align 16
- at __omp_offloading_10302_bd7e0_main_l13_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_10302_bd7e0_main_l16_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_10302_bd7e0_main_l13_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_10302_bd7e0_main_l16_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
;.
@@ -40,8 +40,8 @@ target triple = "nvptx64"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 22, ptr @[[GLOB0]] }, align 8
; CHECK: @[[I_SHARED:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] undef, align 16
; CHECK: @[[I_I_SHARED:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] undef, align 16
-; CHECK: @[[__OMP_OFFLOADING_10302_BD7E0_MAIN_L13_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; CHECK: @[[__OMP_OFFLOADING_10302_BD7E0_MAIN_L16_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[__OMP_OFFLOADING_10302_BD7E0_MAIN_L13_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[__OMP_OFFLOADING_10302_BD7E0_MAIN_L16_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
define weak_odr protected void @__omp_offloading_10302_bd7e0_main_l13(ptr %dyn, i64 noundef %i) local_unnamed_addr "kernel" {
; CHECK-LABEL: @__omp_offloading_10302_bd7e0_main_l13(
diff --git a/llvm/test/Transforms/OpenMP/parallel_level_fold.ll b/llvm/test/Transforms/OpenMP/parallel_level_fold.ll
index 5692197f8093946..d26dd74f0e7f332 100644
--- a/llvm/test/Transforms/OpenMP/parallel_level_fold.ll
+++ b/llvm/test/Transforms/OpenMP/parallel_level_fold.ll
@@ -3,18 +3,18 @@
target triple = "nvptx64"
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@G = external global i16
- at none_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
- at spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
- at parallel_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at none_spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at spmd_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at parallel_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
;.
; CHECK: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i16
-; CHECK: @[[NONE_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
-; CHECK: @[[SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
-; CHECK: @[[PARALLEL_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[NONE_SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[SPMD_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[PARALLEL_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
;.
define weak void @none_spmd() "kernel" {
; CHECK-LABEL: define {{[^@]+}}@none_spmd
diff --git a/llvm/test/Transforms/OpenMP/remove_globalization.ll b/llvm/test/Transforms/OpenMP/remove_globalization.ll
index 0c27e9c86c7fd02..40d9df835b4f7d4 100644
--- a/llvm/test/Transforms/OpenMP/remove_globalization.ll
+++ b/llvm/test/Transforms/OpenMP/remove_globalization.ll
@@ -7,8 +7,8 @@ target triple = "nvptx64"
@S = external local_unnamed_addr global ptr
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
- at kernel_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
+ at kernel_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
; UTC_ARGS: --disable
@@ -22,10 +22,10 @@ target triple = "nvptx64"
; Make it a weak definition so we will apply custom state machine rewriting but can't use the body in the reasoning.
;.
; CHECK: @[[S:[a-zA-Z0-9_$"\\.-]+]] = external local_unnamed_addr global ptr
-; CHECK: @[[KERNEL_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[KERNEL_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
;.
; CHECK-DISABLED: @[[S:[a-zA-Z0-9_$"\\.-]+]] = external local_unnamed_addr global ptr
-; CHECK-DISABLED: @[[KERNEL_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK-DISABLED: @[[KERNEL_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
;.
define weak i32 @__kmpc_target_init(ptr %0, ptr) {
; CHECK-LABEL: define {{[^@]+}}@__kmpc_target_init
diff --git a/llvm/test/Transforms/OpenMP/replace_globalization.ll b/llvm/test/Transforms/OpenMP/replace_globalization.ll
index 4dc10eba524134b..0b88249f2097ae5 100644
--- a/llvm/test/Transforms/OpenMP/replace_globalization.ll
+++ b/llvm/test/Transforms/OpenMP/replace_globalization.ll
@@ -15,14 +15,14 @@ target triple = "nvptx64"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@S = external local_unnamed_addr global ptr
@0 = private unnamed_addr constant [113 x i8] c";llvm/test/Transforms/OpenMP/custom_state_machines_remarks.c;__omp_offloading_2a_d80d3d_test_fallback_l11;11;1;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
- at foo_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at bar_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at baz_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at foo_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at bar_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at baz_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
define dso_local void @foo(ptr %dyn) "kernel" {
@@ -129,9 +129,9 @@ declare void @unknown_no_openmp() "llvm.assume"="omp_no_openmp"
; CHECK: @[[S:[a-zA-Z0-9_$"\\.-]+]] = external local_unnamed_addr global ptr
; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [113 x i8] c"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; CHECK: @[[FOO_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; CHECK: @[[BAR_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; CHECK: @[[BAZ_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[FOO_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[BAR_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[BAZ_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; CHECK: @[[OFFSET:[a-zA-Z0-9_$"\\.-]+]] = global i32 undef
; CHECK: @[[STACK:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [1024 x i8] undef
; CHECK: @[[X_SHARED:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [16 x i8] poison, align 4
diff --git a/llvm/test/Transforms/OpenMP/single_threaded_execution.ll b/llvm/test/Transforms/OpenMP/single_threaded_execution.ll
index 6de658eb63e90ab..c186e5f04f092d4 100644
--- a/llvm/test/Transforms/OpenMP/single_threaded_execution.ll
+++ b/llvm/test/Transforms/OpenMP/single_threaded_execution.ll
@@ -5,11 +5,11 @@
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [1 x i8] c"\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
- at kernel_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at kernel_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
; CHECK-NOT: [openmp-opt] Basic block @kernel entry is executed by a single thread.
diff --git a/llvm/test/Transforms/OpenMP/spmdization.ll b/llvm/test/Transforms/OpenMP/spmdization.ll
index a64be8a22a7f974..635dcdc3c52b106 100644
--- a/llvm/test/Transforms/OpenMP/spmdization.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization.ll
@@ -93,17 +93,17 @@
%struct.kmp_task_t = type { ptr, ptr, i32, %union.kmp_cmplrdata_t, %union.kmp_cmplrdata_t }
%union.kmp_cmplrdata_t = type { ptr }
%struct.anon = type {}
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
- at __omp_offloading_fd02_2044372e_sequential_loop_l5_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_fd02_2044372e_sequential_loop_to_stack_var_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_fd02_2044372e_sequential_loop_to_shared_var_l35_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_fd02_2044372e_sequential_loop_to_shared_var_guarded_l50_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_fd02_2044372e_do_not_spmdize_target_l65_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_fd02_2044372e_do_not_spmdize_task_l74_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_fd02_2044372e_sequential_loop_l5_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_fd02_2044372e_sequential_loop_to_stack_var_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_fd02_2044372e_sequential_loop_to_shared_var_l35_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_fd02_2044372e_sequential_loop_to_shared_var_guarded_l50_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_fd02_2044372e_do_not_spmdize_target_l65_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_fd02_2044372e_do_not_spmdize_task_l74_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
; AMDGPU-DISABLED: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; AMDGPU-DISABLED: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
@@ -138,12 +138,12 @@
;.
; AMDGPU: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; AMDGPU: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; AMDGPU: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 22, ptr @[[GLOB0]] }, align 8
; AMDGPU: @[[X_SHARED:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; AMDGPU: @[[X_SHARED_1:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
@@ -151,12 +151,12 @@
;.
; NVPTX: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; NVPTX: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; NVPTX: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 22, ptr @[[GLOB0]] }, align 8
; NVPTX: @[[X_SHARED:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; NVPTX: @[[X_SHARED1:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
@@ -164,12 +164,12 @@
;.
; AMDGPU-DISABLED1: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; AMDGPU-DISABLED1: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; AMDGPU-DISABLED1: @[[X_SHARED:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; AMDGPU-DISABLED1: @[[X_SHARED_1:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; AMDGPU-DISABLED1: @[[__OMP_OUTLINED__1_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
@@ -180,12 +180,12 @@
;.
; AMDGPU-DISABLED2: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; AMDGPU-DISABLED2: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; AMDGPU-DISABLED2: @[[X_SHARED:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; AMDGPU-DISABLED2: @[[X_SHARED_1:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; AMDGPU-DISABLED2: @[[__OMP_OUTLINED__1_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
@@ -196,12 +196,12 @@
;.
; NVPTX-DISABLED1: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; NVPTX-DISABLED1: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED1: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; NVPTX-DISABLED1: @[[X_SHARED:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; NVPTX-DISABLED1: @[[X_SHARED1:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; NVPTX-DISABLED1: @[[__OMP_OUTLINED__1_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
@@ -212,12 +212,12 @@
;.
; NVPTX-DISABLED2: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; NVPTX-DISABLED2: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_STACK_VAR_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_L35_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_SEQUENTIAL_LOOP_TO_SHARED_VAR_GUARDED_L50_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TARGET_L65_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX-DISABLED2: @[[__OMP_OFFLOADING_FD02_2044372E_DO_NOT_SPMDIZE_TASK_L74_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; NVPTX-DISABLED2: @[[X_SHARED:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; NVPTX-DISABLED2: @[[X_SHARED1:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global [4 x i8] poison, align 4
; NVPTX-DISABLED2: @[[__OMP_OUTLINED__1_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
diff --git a/llvm/test/Transforms/OpenMP/spmdization_assumes.ll b/llvm/test/Transforms/OpenMP/spmdization_assumes.ll
index 2db7c1c42daa5ca..205669572e4118f 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_assumes.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_assumes.ll
@@ -14,18 +14,18 @@ target triple = "nvptx64"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
- at __omp_offloading_fd02_404433c2_main_l5_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_fd02_404433c2_main_l5_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
; Function Attrs: alwaysinline convergent norecurse nounwind
;.
; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; CHECK: @[[__OMP_OFFLOADING_FD02_404433C2_MAIN_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[__OMP_OFFLOADING_FD02_404433C2_MAIN_L5_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; CHECK: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 22, ptr @[[GLOB0]] }, align 8
;.
define weak void @__omp_offloading_fd02_404433c2_main_l5(ptr %dyn, ptr nonnull align 8 dereferenceable(8) %x) local_unnamed_addr #0 {
diff --git a/llvm/test/Transforms/OpenMP/spmdization_constant_prop.ll b/llvm/test/Transforms/OpenMP/spmdization_constant_prop.ll
index 8cf7b6c1c8ee32f..6b15dd4f6b9b18e 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_constant_prop.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_constant_prop.ll
@@ -2,7 +2,7 @@
;
; Verify we change it to SPMD mode but also avoid propagating the old mode (=generic) into the __kmpc_target_init function.
;
-; CHECK: @__omp_offloading_20_11e3950_main_l12_kernel_environment = local_unnamed_addr addrspace(1) constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr addrspacecast (ptr addrspace(1) @1 to ptr), ptr null }
+; CHECK: @__omp_offloading_20_11e3950_main_l12_kernel_environment = local_unnamed_addr addrspace(1) constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr addrspacecast (ptr addrspace(1) @1 to ptr), ptr null }
; CHECK-NOT: store i32 0, ptr addrspace(3) @IsSPMDMode
; CHECK-NOT: store i32 0, ptr addrspace(3) @IsSPMDMode
; CHECK: store i32 1, ptr addrspace(3) @IsSPMDMode
@@ -14,7 +14,7 @@ target triple = "amdgcn-amd-amdhsa"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.DeviceEnvironmentTy = type { i32, i32, i32, i32 }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
%"struct.(anonymous namespace)::SharedMemorySmartStackTy" = type { [512 x i8], [1024 x i8] }
%"struct.(anonymous namespace)::TeamStateTy" = type { %"struct.(anonymous namespace)::ICVStateTy", i32, ptr }
%"struct.(anonymous namespace)::ICVStateTy" = type { i32, i32, i32, i32, i32, i32 }
@@ -27,7 +27,7 @@ target triple = "amdgcn-amd-amdhsa"
@__omp_rtl_debug_kind = weak_odr hidden local_unnamed_addr addrspace(1) constant i32 0
@__omp_rtl_assume_no_thread_state = weak_odr hidden local_unnamed_addr addrspace(1) constant i32 0
@omptarget_device_environment = weak protected addrspace(4) global %struct.DeviceEnvironmentTy undef, align 4
- at __omp_offloading_20_11e3950_main_l12_kernel_environment = local_unnamed_addr addrspace(1) constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr addrspacecast (ptr addrspace(1) @1 to ptr), ptr null }
+ at __omp_offloading_20_11e3950_main_l12_kernel_environment = local_unnamed_addr addrspace(1) constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr addrspacecast (ptr addrspace(1) @1 to ptr), ptr null }
@IsSPMDMode = weak hidden addrspace(3) global i32 undef, align 4
@.str.12 = private unnamed_addr addrspace(4) constant [47 x i8] c"ValueRAII initialization with wrong old value!\00", align 1
@_ZN12_GLOBAL__N_122SharedMemorySmartStackE = internal addrspace(3) global %"struct.(anonymous namespace)::SharedMemorySmartStackTy" undef, align 16
diff --git a/llvm/test/Transforms/OpenMP/spmdization_guarding.ll b/llvm/test/Transforms/OpenMP/spmdization_guarding.ll
index 9c05ab3cc5d5639..c2ca3dc045ac2d9 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_guarding.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_guarding.ll
@@ -35,13 +35,13 @@ target datalayout = "e-i64:64-i128:128-v16:16-v32:32-n16:32:64"
target triple = "nvptx64"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@LocGlob = private unnamed_addr addrspace(5) global i32 43
- at __omp_offloading_2a_fbfa7a_sequential_loop_l6_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_2a_fbfa7a_sequential_loop_l6_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
; Function Attrs: convergent norecurse nounwind
@@ -49,13 +49,13 @@ target triple = "nvptx64"
; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK: @[[LOCGLOB:[a-zA-Z0-9_$"\\.-]+]] = private unnamed_addr addrspace(5) global i32 43
-; CHECK: @[[__OMP_OFFLOADING_2A_FBFA7A_SEQUENTIAL_LOOP_L6_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[__OMP_OFFLOADING_2A_FBFA7A_SEQUENTIAL_LOOP_L6_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; CHECK: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 22, ptr @[[GLOB0]] }, align 8
;.
; CHECK-DISABLED: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK-DISABLED: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK-DISABLED: @[[LOCGLOB:[a-zA-Z0-9_$"\\.-]+]] = private unnamed_addr addrspace(5) global i32 43
-; CHECK-DISABLED: @[[__OMP_OFFLOADING_2A_FBFA7A_SEQUENTIAL_LOOP_L6_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK-DISABLED: @[[__OMP_OFFLOADING_2A_FBFA7A_SEQUENTIAL_LOOP_L6_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; CHECK-DISABLED: @[[__OMP_OUTLINED__1_WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
;.
define weak void @__omp_offloading_2a_fbfa7a_sequential_loop_l6(ptr %dyn, ptr %x, i64 %N) #0 {
diff --git a/llvm/test/Transforms/OpenMP/spmdization_guarding_two_reaching_kernels.ll b/llvm/test/Transforms/OpenMP/spmdization_guarding_two_reaching_kernels.ll
index 00c8642a6034a5d..c15bae1d448175e 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_guarding_two_reaching_kernels.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_guarding_two_reaching_kernels.ll
@@ -30,31 +30,31 @@ target datalayout = "e-i64:64-i128:128-v16:16-v32:32-n16:32:64"
target triple = "nvptx64"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@2 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 2, i32 0, ptr @0 }, align 8
@G = external global i32, align 4
- at __omp_offloading_2b_10393b5_spmd_l12_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_2b_10393b5_generic_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_2b_10393b5_spmd_l12_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_2b_10393b5_generic_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
;.
; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
-; CHECK: @[[__OMP_OFFLOADING_2B_10393B5_SPMD_L12_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; CHECK: @[[__OMP_OFFLOADING_2B_10393B5_GENERIC_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[__OMP_OFFLOADING_2B_10393B5_SPMD_L12_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[__OMP_OFFLOADING_2B_10393B5_GENERIC_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; CHECK: @[[GLOB3:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 22, ptr @[[GLOB0]] }, align 8
;.
; CHECK-DISABLE-SPMDIZATION: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK-DISABLE-SPMDIZATION: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK-DISABLE-SPMDIZATION: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK-DISABLE-SPMDIZATION: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i32, align 4
-; CHECK-DISABLE-SPMDIZATION: @[[__OMP_OFFLOADING_2B_10393B5_SPMD_L12_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; CHECK-DISABLE-SPMDIZATION: @[[__OMP_OFFLOADING_2B_10393B5_GENERIC_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK-DISABLE-SPMDIZATION: @[[__OMP_OFFLOADING_2B_10393B5_SPMD_L12_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK-DISABLE-SPMDIZATION: @[[__OMP_OFFLOADING_2B_10393B5_GENERIC_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
define weak void @__omp_offloading_2b_10393b5_spmd_l12(ptr %dyn) "kernel" #0 {
; CHECK-LABEL: define {{[^@]+}}@__omp_offloading_2b_10393b5_spmd_l12
diff --git a/llvm/test/Transforms/OpenMP/spmdization_indirect.ll b/llvm/test/Transforms/OpenMP/spmdization_indirect.ll
index 8ae94cf9fe37d03..d5e864eea0e78db 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_indirect.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_indirect.ll
@@ -3,30 +3,30 @@
; RUN: opt --mtriple=nvptx64-- -S -passes=openmp-opt < %s | FileCheck %s --check-prefixes=NVPTX
%struct.ident_t = type { i32, i32, i32, i32, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
- at spmd_callees_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at spmd_callees_metadata_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at spmd_and_non_spmd_callees_metadata_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at spmd_and_non_spmd_callee_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at spmd_callees_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at spmd_callees_metadata_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at spmd_and_non_spmd_callees_metadata_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at spmd_and_non_spmd_callee_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
;.
; AMDGPU: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; AMDGPU: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; AMDGPU: @[[SPMD_CALLEES_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[SPMD_CALLEES_METADATA_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[SPMD_AND_NON_SPMD_CALLEES_METADATA_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; AMDGPU: @[[SPMD_AND_NON_SPMD_CALLEE_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[SPMD_CALLEES_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[SPMD_CALLEES_METADATA_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[SPMD_AND_NON_SPMD_CALLEES_METADATA_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; AMDGPU: @[[SPMD_AND_NON_SPMD_CALLEE_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; NVPTX: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; NVPTX: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
-; NVPTX: @[[SPMD_CALLEES_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[SPMD_CALLEES_METADATA_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[SPMD_AND_NON_SPMD_CALLEES_METADATA_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; NVPTX: @[[SPMD_AND_NON_SPMD_CALLEE_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[SPMD_CALLEES_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 1, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[SPMD_CALLEES_METADATA_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[SPMD_AND_NON_SPMD_CALLEES_METADATA_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; NVPTX: @[[SPMD_AND_NON_SPMD_CALLEE_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
define weak void @spmd_callees(i1 %c) #0 {
; AMDGPU-LABEL: define {{[^@]+}}@spmd_callees
diff --git a/llvm/test/Transforms/OpenMP/spmdization_kernel_env_dep.ll b/llvm/test/Transforms/OpenMP/spmdization_kernel_env_dep.ll
index 711d0460085068e..0120cb16cf09a38 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_kernel_env_dep.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_kernel_env_dep.ll
@@ -5,14 +5,14 @@ target datalayout = "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:3
target triple = "amdgcn-amd-amdhsa"
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy.8, ptr, ptr }
-%struct.ConfigurationEnvironmentTy.8 = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy.8 = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@IsSPMDMode = internal addrspace(3) global i32 undef
- at __omp_offloading_10302_b20a40e_main_l4_kernel_environment = addrspace(1) constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy.8 { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr addrspacecast (ptr addrspace(1) null to ptr), ptr addrspacecast (ptr addrspace(1) null to ptr) }
+ at __omp_offloading_10302_b20a40e_main_l4_kernel_environment = addrspace(1) constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy.8 { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr addrspacecast (ptr addrspace(1) null to ptr), ptr addrspacecast (ptr addrspace(1) null to ptr) }
;.
; AMDGPU: @[[ISSPMDMODE:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global i32 undef
-; AMDGPU: @[[__OMP_OFFLOADING_10302_B20A40E_MAIN_L4_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY_8:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr addrspacecast (ptr addrspace(1) null to ptr), ptr addrspacecast (ptr addrspace(1) null to ptr) }
+; AMDGPU: @[[__OMP_OFFLOADING_10302_B20A40E_MAIN_L4_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY_8:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr addrspacecast (ptr addrspace(1) null to ptr), ptr addrspacecast (ptr addrspace(1) null to ptr) }
;.
define i32 @fputs() {
; AMDGPU-LABEL: define {{[^@]+}}@fputs
diff --git a/llvm/test/Transforms/OpenMP/spmdization_no_guarding_two_reaching_kernels.ll b/llvm/test/Transforms/OpenMP/spmdization_no_guarding_two_reaching_kernels.ll
index 3cc7c2559ad2f5f..fdfc1cb08442c7b 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_no_guarding_two_reaching_kernels.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_no_guarding_two_reaching_kernels.ll
@@ -31,30 +31,30 @@ target datalayout = "e-i64:64-i128:128-v16:16-v32:32-n16:32:64"
target triple = "nvptx64"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@2 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 2, i32 0, ptr @0 }, align 8
@G = external addrspace(5) global i32, align 4
- at __omp_offloading_2b_10393b5_spmd_l12_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_2b_10393b5_generic_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_2b_10393b5_spmd_l12_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_2b_10393b5_generic_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
;.
; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external addrspace(5) global i32, align 4
-; CHECK: @[[__OMP_OFFLOADING_2B_10393B5_SPMD_L12_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; CHECK: @[[__OMP_OFFLOADING_2B_10393B5_GENERIC_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[__OMP_OFFLOADING_2B_10393B5_SPMD_L12_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK: @[[__OMP_OFFLOADING_2B_10393B5_GENERIC_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
;.
; CHECK-DISABLE-SPMDIZATION: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK-DISABLE-SPMDIZATION: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK-DISABLE-SPMDIZATION: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 2, i32 0, ptr @[[GLOB0]] }, align 8
; CHECK-DISABLE-SPMDIZATION: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external addrspace(5) global i32, align 4
-; CHECK-DISABLE-SPMDIZATION: @[[__OMP_OFFLOADING_2B_10393B5_SPMD_L12_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
-; CHECK-DISABLE-SPMDIZATION: @[[__OMP_OFFLOADING_2B_10393B5_GENERIC_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK-DISABLE-SPMDIZATION: @[[__OMP_OFFLOADING_2B_10393B5_SPMD_L12_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
+; CHECK-DISABLE-SPMDIZATION: @[[__OMP_OFFLOADING_2B_10393B5_GENERIC_L20_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[GLOB1]], ptr null }
; CHECK-DISABLE-SPMDIZATION: @[[__OMP_OUTLINED___WRAPPER_ID:[a-zA-Z0-9_$"\\.-]+]] = private constant i8 undef
;.
define weak void @__omp_offloading_2b_10393b5_spmd_l12(ptr %dyn) #0 {
diff --git a/llvm/test/Transforms/OpenMP/spmdization_remarks.ll b/llvm/test/Transforms/OpenMP/spmdization_remarks.ll
index e338a8007534949..28df2f52491308f 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_remarks.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_remarks.ll
@@ -38,7 +38,7 @@ target triple = "nvptx64"
;; }
%struct.ident_t = type { i32, i32, i32, i32, ptr }
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
@0 = private unnamed_addr constant [103 x i8] c";llvm/test/Transforms/OpenMP/spmdization_remarks.c;__omp_offloading_2a_d80d3d_test_fallback_l11;11;1;;\00", align 1
@@ -57,8 +57,8 @@ target triple = "nvptx64"
@13 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 2, i32 0, ptr @12 }, align 8
@G = external global i32
- at __omp_offloading_2a_d80d3d_test_fallback_l11_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
- at __omp_offloading_2a_d80d3d_test_no_fallback_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_2a_d80d3d_test_fallback_l11_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
+ at __omp_offloading_2a_d80d3d_test_no_fallback_l20_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
; Function Attrs: convergent norecurse nounwind
diff --git a/llvm/test/Transforms/OpenMP/value-simplify-openmp-opt.ll b/llvm/test/Transforms/OpenMP/value-simplify-openmp-opt.ll
index f322e745c714b60..733d2906fbee3b7 100644
--- a/llvm/test/Transforms/OpenMP/value-simplify-openmp-opt.ll
+++ b/llvm/test/Transforms/OpenMP/value-simplify-openmp-opt.ll
@@ -4,7 +4,7 @@
target triple = "amdgcn-amd-amdhsa"
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
@G = internal addrspace(3) global i32 undef, align 4
@@ -27,7 +27,7 @@ target triple = "amdgcn-amd-amdhsa"
@UAA3 = internal addrspace(3) global i32 undef, align 4
@UANA1 = internal addrspace(3) global i32 undef, align 4
@str = private unnamed_addr addrspace(4) constant [1 x i8] c"\00", align 1
- at kernel_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at kernel_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
; Make sure we do not delete the stores to @G without also replacing the load with `1`.
;.
@@ -51,7 +51,7 @@ target triple = "amdgcn-amd-amdhsa"
; CHECK: @[[UAA3:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global i32 undef, align 4
; CHECK: @[[UANA1:[a-zA-Z0-9_$"\\.-]+]] = internal addrspace(3) global i32 undef, align 4
; CHECK: @[[STR:[a-zA-Z0-9_$"\\.-]+]] = private unnamed_addr addrspace(4) constant [1 x i8] zeroinitializer, align 1
-; CHECK: @[[KERNEL_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+; CHECK: @[[KERNEL_KERNEL_ENVIRONMENT:[a-zA-Z0-9_$"\\.-]+]] = local_unnamed_addr constant [[STRUCT_KERNELENVIRONMENTTY:%.*]] { [[STRUCT_CONFIGURATIONENVIRONMENTTY:%.*]] { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
;.
define void @kernel(ptr %dyn) "kernel" {
;
diff --git a/llvm/test/Transforms/PhaseOrdering/openmp-opt-module.ll b/llvm/test/Transforms/PhaseOrdering/openmp-opt-module.ll
index a64107b25d520b3..0faa63cfa972e6e 100644
--- a/llvm/test/Transforms/PhaseOrdering/openmp-opt-module.ll
+++ b/llvm/test/Transforms/PhaseOrdering/openmp-opt-module.ll
@@ -1,12 +1,12 @@
; RUN: opt -passes='default<O2>' -pass-remarks-missed=openmp-opt < %s 2>&1 | FileCheck %s --check-prefix=MODULE
target datalayout = "e-i64:64-i128:128-v16:16-v32:32-n16:32:64"
-%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32 }
+%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
@.str = private unnamed_addr constant [13 x i8] c"Alloc Shared\00", align 1
@S = external local_unnamed_addr global ptr
- at foo_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
+ at foo_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr null, ptr null }
; MODULE: remark: openmp_opt_module.c:5:7: Found thread data sharing on the GPU. Expect degraded performance due to data globalization.
diff --git a/mlir/test/Target/LLVMIR/omptarget-region-device-llvm.mlir b/mlir/test/Target/LLVMIR/omptarget-region-device-llvm.mlir
index 78bab6ece73e6b0..f537eb5c3f572b4 100644
--- a/mlir/test/Target/LLVMIR/omptarget-region-device-llvm.mlir
+++ b/mlir/test/Target/LLVMIR/omptarget-region-device-llvm.mlir
@@ -30,7 +30,7 @@ module attributes {omp.is_target_device = true} {
// CHECK: @[[SRC_LOC:.*]] = private unnamed_addr constant [23 x i8] c"{{[^"]*}}", align 1
// CHECK: @[[IDENT:.*]] = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 22, ptr @[[SRC_LOC]] }, align 8
// CHECK: @[[DYNA_ENV:.*]] = weak_odr protected global %struct.DynamicEnvironmentTy zeroinitializer
-// CHECK: @[[KERNEL_ENV:.*]] = weak_odr protected constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[IDENT]], ptr @[[DYNA_ENV]] }
+// CHECK: @[[KERNEL_ENV:.*]] = weak_odr protected constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 1, i8 1, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @[[IDENT]], ptr @[[DYNA_ENV]] }
// CHECK: define weak_odr protected void @__omp_offloading_{{[^_]+}}_{{[^_]+}}_omp_target_region__l{{[0-9]+}}(ptr %[[DYN_PTR:.*]], ptr %[[ADDR_A:.*]], ptr %[[ADDR_B:.*]], ptr %[[ADDR_C:.*]])
// CHECK: %[[TMP_A:.*]] = alloca ptr, align 8
// CHECK: store ptr %[[ADDR_A]], ptr %[[TMP_A]], align 8
diff --git a/openmp/libomptarget/DeviceRTL/include/Interface.h b/openmp/libomptarget/DeviceRTL/include/Interface.h
index 24de620759c4194..54c3b3aea93cb03 100644
--- a/openmp/libomptarget/DeviceRTL/include/Interface.h
+++ b/openmp/libomptarget/DeviceRTL/include/Interface.h
@@ -232,15 +232,17 @@ void __kmpc_target_deinit();
///{
void *__kmpc_reduction_get_fixed_buffer();
-int32_t __kmpc_nvptx_parallel_reduce_nowait_v2(
- IdentTy *Loc, int32_t TId, int32_t num_vars, uint64_t reduce_size,
- void *reduce_data, ShuffleReductFnTy shflFct, InterWarpCopyFnTy cpyFct);
+int32_t __kmpc_nvptx_parallel_reduce_nowait_v2(IdentTy *Loc,
+ uint64_t reduce_data_size,
+ void *reduce_data,
+ ShuffleReductFnTy shflFct,
+ InterWarpCopyFnTy cpyFct);
int32_t __kmpc_nvptx_teams_reduce_nowait_v2(
- IdentTy *Loc, int32_t TId, void *GlobalBuffer, uint32_t num_of_records,
- void *reduce_data, ShuffleReductFnTy shflFct, InterWarpCopyFnTy cpyFct,
- ListGlobalFnTy lgcpyFct, ListGlobalFnTy lgredFct, ListGlobalFnTy glcpyFct,
- ListGlobalFnTy glredFct);
+ IdentTy *Loc, void *GlobalBuffer, uint32_t num_of_records,
+ uint64_t reduce_data_size, void *reduce_data, ShuffleReductFnTy shflFct,
+ InterWarpCopyFnTy cpyFct, ListGlobalFnTy lgcpyFct, ListGlobalFnTy lgredFct,
+ ListGlobalFnTy glcpyFct, ListGlobalFnTy glredFct);
///}
/// Synchronization
diff --git a/openmp/libomptarget/DeviceRTL/src/Reduction.cpp b/openmp/libomptarget/DeviceRTL/src/Reduction.cpp
index fc5a3b08cb9d07f..41e1dd5180613f5 100644
--- a/openmp/libomptarget/DeviceRTL/src/Reduction.cpp
+++ b/openmp/libomptarget/DeviceRTL/src/Reduction.cpp
@@ -65,12 +65,10 @@ static uint32_t gpu_irregular_simd_reduce(void *reduce_data,
}
#endif
-static int32_t nvptx_parallel_reduce_nowait(int32_t TId, int32_t num_vars,
- uint64_t reduce_size,
+static int32_t nvptx_parallel_reduce_nowait(
void *reduce_data,
ShuffleReductFnTy shflFct,
- InterWarpCopyFnTy cpyFct,
- bool isSPMDExecutionMode, bool) {
+ InterWarpCopyFnTy cpyFct) {
uint32_t BlockThreadId = mapping::getThreadIdInBlock();
if (mapping::isMainThreadInGenericMode(/* IsSPMD */ false))
BlockThreadId = 0;
@@ -155,7 +153,7 @@ static int32_t nvptx_parallel_reduce_nowait(int32_t TId, int32_t num_vars,
// Get the OMP thread Id. This is
diff erent from BlockThreadId in the case of
// an L2 parallel region.
- return TId == 0;
+ return BlockThreadId == 0;
#endif // __CUDA_ARCH__ >= 700
}
@@ -170,19 +168,19 @@ uint32_t kmpcMin(uint32_t x, uint32_t y) { return x < y ? x : y; }
} // namespace
extern "C" {
-int32_t __kmpc_nvptx_parallel_reduce_nowait_v2(
- IdentTy *Loc, int32_t TId, int32_t num_vars, uint64_t reduce_size,
- void *reduce_data, ShuffleReductFnTy shflFct, InterWarpCopyFnTy cpyFct) {
- return nvptx_parallel_reduce_nowait(TId, num_vars, reduce_size, reduce_data,
- shflFct, cpyFct, mapping::isSPMDMode(),
- false);
+int32_t __kmpc_nvptx_parallel_reduce_nowait_v2(IdentTy *Loc,
+ uint64_t reduce_data_size,
+ void *reduce_data,
+ ShuffleReductFnTy shflFct,
+ InterWarpCopyFnTy cpyFct) {
+ return nvptx_parallel_reduce_nowait(reduce_data, shflFct, cpyFct);
}
/// Mostly like _v2 but with the builtin assumption that we have less than
/// num_of_records (by default 1024) teams.
int32_t __kmpc_nvptx_teams_reduce_nowait_v3(
- IdentTy *Loc, int32_t TId, void *__restrict__ GlobalBuffer,
- uint32_t num_of_records, void *reduce_data, ShuffleReductFnTy shflFct,
+ IdentTy *Loc, void *__restrict__ GlobalBuffer, uint32_t num_of_records,
+ uint64_t reduce_data_size, void *reduce_data, ShuffleReductFnTy shflFct,
InterWarpCopyFnTy cpyFct, ListGlobalFnTy lgcpyFct, ListGlobalFnTy lgredFct,
ListGlobalFnTy glcpyFct, ListGlobalFnTy glredFct) {
// Terminate all threads in non-SPMD mode except for the main thread.
@@ -267,10 +265,10 @@ int32_t __kmpc_nvptx_teams_reduce_nowait_v3(
}
int32_t __kmpc_nvptx_teams_reduce_nowait_v2(
- IdentTy *Loc, int32_t TId, void *GlobalBuffer, uint32_t num_of_records,
- void *reduce_data, ShuffleReductFnTy shflFct, InterWarpCopyFnTy cpyFct,
- ListGlobalFnTy lgcpyFct, ListGlobalFnTy lgredFct, ListGlobalFnTy glcpyFct,
- ListGlobalFnTy glredFct) {
+ IdentTy *Loc, void *GlobalBuffer, uint32_t num_of_records,
+ uint64_t reduce_data_size, void *reduce_data, ShuffleReductFnTy shflFct,
+ InterWarpCopyFnTy cpyFct, ListGlobalFnTy lgcpyFct, ListGlobalFnTy lgredFct,
+ ListGlobalFnTy glcpyFct, ListGlobalFnTy glredFct) {
// The first check is a compile time constant, the second one a runtime check.
// If the first one succeeds we will use the specialized version.
if ((state::getKernelEnvironment().Configuration.MaxTeams >= 0 &&
@@ -278,8 +276,8 @@ int32_t __kmpc_nvptx_teams_reduce_nowait_v2(
num_of_records == 1024) ||
(omp_get_num_teams() <= num_of_records))
return __kmpc_nvptx_teams_reduce_nowait_v3(
- Loc, TId, GlobalBuffer, num_of_records, reduce_data, shflFct, cpyFct,
- lgcpyFct, lgredFct, glcpyFct, glredFct);
+ Loc, GlobalBuffer, num_of_records, reduce_data_size, reduce_data,
+ shflFct, cpyFct, lgcpyFct, lgredFct, glcpyFct, glredFct);
// Terminate all threads in non-SPMD mode except for the master thread.
uint32_t ThreadId = mapping::getThreadIdInBlock();
diff --git a/openmp/libomptarget/include/Environment.h b/openmp/libomptarget/include/Environment.h
index bd493e8a0be78f1..3f3dce692f788db 100644
--- a/openmp/libomptarget/include/Environment.h
+++ b/openmp/libomptarget/include/Environment.h
@@ -86,7 +86,8 @@ struct ConfigurationEnvironmentTy {
int32_t MaxThreads = -1;
int32_t MinTeams = -1;
int32_t MaxTeams = -1;
- int32_t ReductionBufferSize = 0;
+ int32_t ReductionDataSize = 0;
+ int32_t ReductionBufferLength = 0;
//}
};
diff --git a/openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp b/openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp
index 106e7a68cd3ae3c..a47481104430f36 100644
--- a/openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp
+++ b/openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp
@@ -402,7 +402,7 @@ Error GenericKernelTy::init(GenericDeviceTy &GenericDevice,
DP("Failed to read kernel environment for '%s': %s\n"
"Using default SPMD (2) execution mode\n",
Name, ErrStr.data());
- assert(KernelEnvironment.Configuration.ReductionBufferSize == 0 &&
+ assert(KernelEnvironment.Configuration.ReductionDataSize == 0 &&
"Default initialization failed.");
}
@@ -440,9 +440,11 @@ GenericKernelTy::getKernelLaunchEnvironment(
/// async data transfer.
auto &LocalKLE = (*AsyncInfoWrapper).KernelLaunchEnvironment;
LocalKLE = KernelLaunchEnvironment;
- if (KernelEnvironment.Configuration.ReductionBufferSize) {
+ if (KernelEnvironment.Configuration.ReductionDataSize &&
+ KernelEnvironment.Configuration.ReductionBufferLength) {
auto AllocOrErr = GenericDevice.dataAlloc(
- KernelEnvironment.Configuration.ReductionBufferSize,
+ KernelEnvironment.Configuration.ReductionDataSize *
+ KernelEnvironment.Configuration.ReductionBufferLength,
/*HostPtr=*/nullptr, TargetAllocTy::TARGET_ALLOC_DEVICE);
if (!AllocOrErr)
return AllocOrErr.takeError();
More information about the cfe-commits
mailing list