[PATCH] D89746: [OpenMP] Emit calls to int64_t functions for amdgcn

[Jon Chesterfield via Phabricator via llvm-commits]

JonChesterfield created this revision.
JonChesterfield added a reviewer: jdoerfert.
Herald added subscribers: llvm-commits, guansong, hiraditya, yaxunl.
Herald added a project: LLVM.
JonChesterfield requested review of this revision.
Herald added a subscriber: sstefan1.

[OpenMP] Emit calls to int64_t functions for amdgcn

Two functions, syncwarp and active_thread_mask, return lanemask_t. Currently
this is assumed to be int32, which is true for nvptx. Patch makes the type
target architecture dependent.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D89746

Files:
  llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
  llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
  llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
  llvm/lib/Transforms/IPO/OpenMPOpt.cpp


Index: llvm/lib/Transforms/IPO/OpenMPOpt.cpp
===================================================================
--- llvm/lib/Transforms/IPO/OpenMPOpt.cpp
+++ llvm/lib/Transforms/IPO/OpenMPOpt.cpp
@@ -331,6 +331,7 @@
   /// in OpenMPKinds.def.
   void initializeRuntimeFunctions() {
     Module &M = *((*ModuleSlice.begin())->getParent());
+    Type *LanemaskTy = OMPBuilder.getLanemaskType();
 
     // Helper macros for handling __VA_ARGS__ in OMP_RTL
 #define OMP_TYPE(VarName, ...)                                                 \
Index: llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
===================================================================
--- llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
+++ llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
@@ -16,6 +16,7 @@
 
 #include "llvm/ADT/StringRef.h"
 #include "llvm/ADT/StringSwitch.h"
+#include "llvm/ADT/Triple.h"
 #include "llvm/IR/CFG.h"
 #include "llvm/IR/DebugInfo.h"
 #include "llvm/IR/IRBuilder.h"
@@ -62,6 +63,7 @@
 OpenMPIRBuilder::getOrCreateRuntimeFunction(Module &M, RuntimeFunction FnID) {
   FunctionType *FnTy = nullptr;
   Function *Fn = nullptr;
+  Type *LanemaskTy = getLanemaskType();
 
   // Try to find the declation in the module first.
   switch (FnID) {
@@ -217,6 +219,14 @@
   return Ident;
 }
 
+Type *OpenMPIRBuilder::getLanemaskType() {
+  LLVMContext &Ctx = M.getContext();
+  Triple triple(M.getTargetTriple());
+
+  // This test is adequate until deviceRTL has finer grained lane widths
+  return triple.isAMDGCN() ? Type::getInt64Ty(Ctx) : Type::getInt32Ty(Ctx);
+}
+
 Constant *OpenMPIRBuilder::getOrCreateSrcLocStr(StringRef LocStr) {
   Constant *&SrcLocStr = SrcLocStrMap[LocStr];
   if (!SrcLocStr) {
@@ -1177,6 +1187,8 @@
 void OpenMPIRBuilder::initializeTypes(Module &M) {
   LLVMContext &Ctx = M.getContext();
   StructType *T;
+  LanemaskTy = getLanemaskType();
+
 #define OMP_TYPE(VarName, InitValue) VarName = InitValue;
 #define OMP_ARRAY_TYPE(VarName, ElemTy, ArraySize)                             \
   VarName##Ty = ArrayType::get(ElemTy, ArraySize);                             \
Index: llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
===================================================================
--- llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
+++ llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
@@ -553,8 +553,9 @@
           Int16, VoidPtrPtr)
 __OMP_RTL(__kmpc_restore_team_static_memory, false, Void, Int16, Int16)
 __OMP_RTL(__kmpc_barrier_simple_spmd, false, Void, IdentPtr, Int32)
-__OMP_RTL(__kmpc_warp_active_thread_mask, false, Int32, )
-__OMP_RTL(__kmpc_syncwarp, false, Void, Int32)
+
+__OMP_RTL(__kmpc_warp_active_thread_mask, false, LanemaskTy,)
+__OMP_RTL(__kmpc_syncwarp, false, Void, LanemaskTy)
 
 __OMP_RTL(__last, false, Void, )
 
Index: llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
===================================================================
--- llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
+++ llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
@@ -226,6 +226,9 @@
                           omp::IdentFlag Flags = omp::IdentFlag(0),
                           unsigned Reserve2Flags = 0);
 
+  // Get the type corresponding to __kmpc_impl_lanemask_t from the deviceRTL
+  Type *getLanemaskType();
+
   /// Generate control flow and cleanup for cancellation.
   ///
   /// \param CancelFlag Flag indicating if the cancellation is performed.
@@ -427,6 +430,7 @@
   /// values.
   ///
   ///{
+  Type *LanemaskTy = nullptr;
 #define OMP_TYPE(VarName, InitValue) Type *VarName = nullptr;
 #define OMP_ARRAY_TYPE(VarName, ElemTy, ArraySize)                             \
   ArrayType *VarName##Ty = nullptr;                                            \


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D89746.299192.patch
Type: text/x-patch
Size: 3690 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20201019/0228bb29/attachment.bin>