[Mlir-commits] [flang] [llvm] [mlir] [flang][OpenMP][MLIR] Add MLIR op for loop directive (PR #113911)
Kareem Ergawy
llvmlistbot at llvm.org
Mon Oct 28 07:09:34 PDT 2024
https://github.com/ergawy updated https://github.com/llvm/llvm-project/pull/113911
>From 6e083978e943208bf0150340be5d720375712ea3 Mon Sep 17 00:00:00 2001
From: ergawy <kareem.ergawy at amd.com>
Date: Fri, 25 Oct 2024 04:00:52 -0500
Subject: [PATCH 1/2] [flang][OpenMP] Parase `bind` clause for `loop`
direcitve.
Adds parsing for the `bind` clause. The clause was already part of the
`loop` direcitve's definition but parsing was still missing.
---
flang/include/flang/Parser/dump-parse-tree.h | 2 ++
flang/include/flang/Parser/parse-tree.h | 7 +++++++
flang/lib/Parser/openmp-parsers.cpp | 8 ++++++++
flang/test/Parser/OpenMP/target-loop-unparse.f90 | 16 ++++++++++++++--
llvm/include/llvm/Frontend/OpenMP/OMP.td | 1 +
5 files changed, 32 insertions(+), 2 deletions(-)
diff --git a/flang/include/flang/Parser/dump-parse-tree.h b/flang/include/flang/Parser/dump-parse-tree.h
index ccdfe980f6f38c..ea02ae27347a1c 100644
--- a/flang/include/flang/Parser/dump-parse-tree.h
+++ b/flang/include/flang/Parser/dump-parse-tree.h
@@ -551,6 +551,8 @@ class ParseTreeDumper {
NODE_ENUM(OmpGrainsizeClause, Prescriptiveness)
NODE(parser, OmpNumTasksClause)
NODE_ENUM(OmpNumTasksClause, Prescriptiveness)
+ NODE(parser, OmpBindClause)
+ NODE_ENUM(OmpBindClause, Type)
NODE(parser, OmpProcBindClause)
NODE_ENUM(OmpProcBindClause, Type)
NODE_ENUM(OmpReductionClause, ReductionModifier)
diff --git a/flang/include/flang/Parser/parse-tree.h b/flang/include/flang/Parser/parse-tree.h
index 2a312e29a3a44d..a08e269003aaab 100644
--- a/flang/include/flang/Parser/parse-tree.h
+++ b/flang/include/flang/Parser/parse-tree.h
@@ -3702,6 +3702,13 @@ struct OmpNumTasksClause {
std::tuple<std::optional<Prescriptiveness>, ScalarIntExpr> t;
};
+// OMP 5.2 11.7.1 bind-clause ->
+// BIND( PARALLEL | TEAMS | THREAD )
+struct OmpBindClause {
+ ENUM_CLASS(Type, Parallel, Teams, Thread)
+ WRAPPER_CLASS_BOILERPLATE(OmpBindClause, Type);
+};
+
// OpenMP Clauses
struct OmpClause {
UNION_CLASS_BOILERPLATE(OmpClause);
diff --git a/flang/lib/Parser/openmp-parsers.cpp b/flang/lib/Parser/openmp-parsers.cpp
index ae0c351fed56d1..06520d3fc2a56c 100644
--- a/flang/lib/Parser/openmp-parsers.cpp
+++ b/flang/lib/Parser/openmp-parsers.cpp
@@ -427,6 +427,12 @@ TYPE_PARSER(construct<OmpLastprivateClause>(
pure(OmpLastprivateClause::LastprivateModifier::Conditional) / ":"),
Parser<OmpObjectList>{}))
+// OMP 5.2 11.7.1 BIND ( PARALLEL | TEAMS | THREAD )
+TYPE_PARSER(construct<OmpBindClause>(
+ "PARALLEL" >> pure(OmpBindClause::Type::Parallel) ||
+ "TEAMS" >> pure(OmpBindClause::Type::Teams) ||
+ "THREAD" >> pure(OmpBindClause::Type::Thread)))
+
TYPE_PARSER(
"ACQUIRE" >> construct<OmpClause>(construct<OmpClause::Acquire>()) ||
"ACQ_REL" >> construct<OmpClause>(construct<OmpClause::AcqRel>()) ||
@@ -441,6 +447,8 @@ TYPE_PARSER(
"ATOMIC_DEFAULT_MEM_ORDER" >>
construct<OmpClause>(construct<OmpClause::AtomicDefaultMemOrder>(
parenthesized(Parser<OmpAtomicDefaultMemOrderClause>{}))) ||
+ "BIND" >> construct<OmpClause>(construct<OmpClause::Bind>(
+ parenthesized(Parser<OmpBindClause>{}))) ||
"COLLAPSE" >> construct<OmpClause>(construct<OmpClause::Collapse>(
parenthesized(scalarIntConstantExpr))) ||
"COPYIN" >> construct<OmpClause>(construct<OmpClause::Copyin>(
diff --git a/flang/test/Parser/OpenMP/target-loop-unparse.f90 b/flang/test/Parser/OpenMP/target-loop-unparse.f90
index 3ee2fcef075a37..b2047070496527 100644
--- a/flang/test/Parser/OpenMP/target-loop-unparse.f90
+++ b/flang/test/Parser/OpenMP/target-loop-unparse.f90
@@ -1,6 +1,8 @@
+! RUN: %flang_fc1 -fdebug-unparse -fopenmp -fopenmp-version=50 %s | \
+! RUN: FileCheck --ignore-case %s
-! RUN: %flang_fc1 -fdebug-unparse -fopenmp %s | FileCheck --ignore-case %s
-! RUN: %flang_fc1 -fdebug-dump-parse-tree -fopenmp %s | FileCheck --check-prefix="PARSE-TREE" %s
+! RUN: %flang_fc1 -fdebug-dump-parse-tree -fopenmp -fopenmp-version=50 %s | \
+! RUN: FileCheck --check-prefix="PARSE-TREE" %s
! Check for parsing of loop directive
@@ -14,6 +16,16 @@ subroutine test_loop
j = j + 1
end do
!$omp end loop
+
+ !PARSE-TREE: OmpBeginLoopDirective
+ !PARSE-TREE-NEXT: OmpLoopDirective -> llvm::omp::Directive = loop
+ !PARSE-TREE-NEXT: OmpClauseList -> OmpClause -> Bind -> OmpBindClause -> Type = Thread
+ !CHECK: !$omp loop
+ !$omp loop bind(thread)
+ do i=1,10
+ j = j + 1
+ end do
+ !$omp end loop
end subroutine
subroutine test_target_loop
diff --git a/llvm/include/llvm/Frontend/OpenMP/OMP.td b/llvm/include/llvm/Frontend/OpenMP/OMP.td
index 70179bab475779..feb8eb5c2abf49 100644
--- a/llvm/include/llvm/Frontend/OpenMP/OMP.td
+++ b/llvm/include/llvm/Frontend/OpenMP/OMP.td
@@ -73,6 +73,7 @@ def OMPC_AtomicDefaultMemOrder : Clause<"atomic_default_mem_order"> {
}
def OMPC_Bind : Clause<"bind"> {
let clangClass = "OMPBindClause";
+ let flangClass = "OmpBindClause";
}
def OMP_CANCELLATION_CONSTRUCT_Parallel : ClauseVal<"parallel", 1, 1> {}
def OMP_CANCELLATION_CONSTRUCT_Loop : ClauseVal<"loop", 2, 1> {}
>From 21371560fcf5cd83160fa8f1199993e6b7e2ac15 Mon Sep 17 00:00:00 2001
From: ergawy <kareem.ergawy at amd.com>
Date: Mon, 28 Oct 2024 00:04:15 -0500
Subject: [PATCH 2/2] [flang][OpenMP][MLIR] Add MLIR op for `loop` directive
Adds MLIR op that corresponds to the `loop` directive.
---
llvm/include/llvm/Frontend/OpenMP/OMP.td | 11 +++++
.../mlir/Dialect/OpenMP/OpenMPClauses.td | 25 +++++++++++
mlir/include/mlir/Dialect/OpenMP/OpenMPOps.td | 44 +++++++++++++++++++
mlir/lib/Dialect/OpenMP/IR/OpenMPDialect.cpp | 23 ++++++++++
mlir/test/Dialect/OpenMP/invalid.mlir | 30 +++++++++++++
mlir/test/Dialect/OpenMP/ops.mlir | 40 +++++++++++++++++
6 files changed, 173 insertions(+)
diff --git a/llvm/include/llvm/Frontend/OpenMP/OMP.td b/llvm/include/llvm/Frontend/OpenMP/OMP.td
index feb8eb5c2abf49..9668ab1e90ecf3 100644
--- a/llvm/include/llvm/Frontend/OpenMP/OMP.td
+++ b/llvm/include/llvm/Frontend/OpenMP/OMP.td
@@ -71,10 +71,21 @@ def OMPC_AtomicDefaultMemOrder : Clause<"atomic_default_mem_order"> {
let clangClass = "OMPAtomicDefaultMemOrderClause";
let flangClass = "OmpAtomicDefaultMemOrderClause";
}
+
+def OMP_BIND_parallel : ClauseVal<"parallel",1,1> {}
+def OMP_BIND_teams : ClauseVal<"teams",2,1> {}
+def OMP_BIND_thread : ClauseVal<"thread",3,1> { let isDefault = true; }
def OMPC_Bind : Clause<"bind"> {
let clangClass = "OMPBindClause";
let flangClass = "OmpBindClause";
+ let enumClauseValue = "BindKind";
+ let allowedClauseValues = [
+ OMP_BIND_parallel,
+ OMP_BIND_teams,
+ OMP_BIND_thread
+ ];
}
+
def OMP_CANCELLATION_CONSTRUCT_Parallel : ClauseVal<"parallel", 1, 1> {}
def OMP_CANCELLATION_CONSTRUCT_Loop : ClauseVal<"loop", 2, 1> {}
def OMP_CANCELLATION_CONSTRUCT_Sections : ClauseVal<"sections", 3, 1> {}
diff --git a/mlir/include/mlir/Dialect/OpenMP/OpenMPClauses.td b/mlir/include/mlir/Dialect/OpenMP/OpenMPClauses.td
index 886554f66afffc..855deab94b2f16 100644
--- a/mlir/include/mlir/Dialect/OpenMP/OpenMPClauses.td
+++ b/mlir/include/mlir/Dialect/OpenMP/OpenMPClauses.td
@@ -107,6 +107,31 @@ class OpenMP_CancelDirectiveNameClauseSkip<
def OpenMP_CancelDirectiveNameClause : OpenMP_CancelDirectiveNameClauseSkip<>;
+//===----------------------------------------------------------------------===//
+// V5.2: [11.7.1] `bind` clause
+//===----------------------------------------------------------------------===//
+
+class OpenMP_BindClauseSkip<
+ bit traits = false, bit arguments = false, bit assemblyFormat = false,
+ bit description = false, bit extraClassDeclaration = false
+ > : OpenMP_Clause<traits, arguments, assemblyFormat, description,
+ extraClassDeclaration> {
+ let arguments = (ins
+ OptionalAttr<BindKindAttr>:$bind_kind
+ );
+
+ let optAssemblyFormat = [{
+ `bind` `(` custom<ClauseAttr>($bind_kind) `)`
+ }];
+
+ let description = [{
+ The `bind` clause specifies the binding region of the construct on which it
+ appears.
+ }];
+}
+
+def OpenMP_BindClause : OpenMP_BindClauseSkip<>;
+
//===----------------------------------------------------------------------===//
// V5.2: [5.7.2] `copyprivate` clause
//===----------------------------------------------------------------------===//
diff --git a/mlir/include/mlir/Dialect/OpenMP/OpenMPOps.td b/mlir/include/mlir/Dialect/OpenMP/OpenMPOps.td
index 626539cb7bde42..4f16108a7a585f 100644
--- a/mlir/include/mlir/Dialect/OpenMP/OpenMPOps.td
+++ b/mlir/include/mlir/Dialect/OpenMP/OpenMPOps.td
@@ -382,6 +382,50 @@ def LoopNestOp : OpenMP_Op<"loop_nest", traits = [
// 2.9.2 Workshare Loop Construct
//===----------------------------------------------------------------------===//
+def LoopOp : OpenMP_Op<"loop", traits = [
+ DeclareOpInterfaceMethods<LoopWrapperInterface>, NoTerminator, SingleBlock,
+ AttrSizedOperandSegments
+ ], clauses = [
+ OpenMP_BindClause, OpenMP_PrivateClause, OpenMP_OrderClause,
+ OpenMP_ReductionClause
+ ], singleRegion = true> {
+ let summary = "loop construct";
+ let description = [{
+ A loop construct specifies that the logical iterations of the associated loops
+ may execute concurrently and permits the encountering threads to execute the
+ loop accordingly. A loop construct is a worksharing construct if its binding
+ region is the innermost enclosing parallel region. Otherwise it is not a work-
+ sharing region. The directive asserts that the iterations of the associated
+ loops may execute in any order, including concurrently. Each logical iteration
+ is executed once per instance of the loop region that is encountered by exactly
+ one thread that is a member of the binding thread set.
+
+ The body region can only contain a single block which must contain a single
+ operation, this operation must be an `omp.loop_nest`.
+
+ ```
+ omp.loop <clauses> {
+ omp.loop_nest (%i1, %i2) : index = (%c0, %c0) to (%c10, %c10) step (%c1, %c1) {
+ %a = load %arrA[%i1, %i2] : memref<?x?xf32>
+ %b = load %arrB[%i1, %i2] : memref<?x?xf32>
+ %sum = arith.addf %a, %b : f32
+ store %sum, %arrC[%i1, %i2] : memref<?x?xf32>
+ omp.yield
+ }
+ }
+ ```
+ }] # clausesDescription;
+
+ let assemblyFormat = clausesAssemblyFormat # [{
+ custom<PrivateReductionRegion>($region, $private_vars, type($private_vars),
+ $private_syms, $reduction_vars, type($reduction_vars), $reduction_byref,
+ $reduction_syms) attr-dict
+ }];
+
+ let hasVerifier = 1;
+ let hasRegionVerifier = 1;
+}
+
def WsloopOp : OpenMP_Op<"wsloop", traits = [
AttrSizedOperandSegments, DeclareOpInterfaceMethods<ComposableOpInterface>,
DeclareOpInterfaceMethods<LoopWrapperInterface>, NoTerminator,
diff --git a/mlir/lib/Dialect/OpenMP/IR/OpenMPDialect.cpp b/mlir/lib/Dialect/OpenMP/IR/OpenMPDialect.cpp
index e1df647d6a3c71..39accbadc8e276 100644
--- a/mlir/lib/Dialect/OpenMP/IR/OpenMPDialect.cpp
+++ b/mlir/lib/Dialect/OpenMP/IR/OpenMPDialect.cpp
@@ -1948,6 +1948,29 @@ LogicalResult LoopWrapperInterface::verifyImpl() {
return success();
}
+//===----------------------------------------------------------------------===//
+// LoopOp
+//===----------------------------------------------------------------------===//
+
+LogicalResult LoopOp::verify() {
+ return verifyReductionVarList(*this, getReductionSyms(), getReductionVars(),
+ getReductionByref());
+}
+
+LogicalResult LoopOp::verifyRegions() {
+ Region ®ion = getRegion();
+
+ // Minimal amount of checks to verify the only nested op is an
+ // `omp.loop_nest`. A more extensive vierfication is done by the
+ // `LoopWrapperInterface` trait but the difference is that `omp.loop` cannot
+ // have another nested `LoopWrapperInterface`.
+ if (range_size(region.getOps()) != 1 || !isa<LoopNestOp>(*region.op_begin()))
+ return emitError() << "`omp.loop` expected to have a single nested "
+ "operation which is a `omp.loop_nest`";
+
+ return success();
+}
+
//===----------------------------------------------------------------------===//
// WsloopOp
//===----------------------------------------------------------------------===//
diff --git a/mlir/test/Dialect/OpenMP/invalid.mlir b/mlir/test/Dialect/OpenMP/invalid.mlir
index fd89ec31c64a60..194daf8c2edf34 100644
--- a/mlir/test/Dialect/OpenMP/invalid.mlir
+++ b/mlir/test/Dialect/OpenMP/invalid.mlir
@@ -2577,3 +2577,33 @@ func.func @omp_taskloop_invalid_composite(%lb: index, %ub: index, %step: index)
} {omp.composite}
return
}
+
+// -----
+
+func.func @omp_loop_invalid_nesting(%lb : index, %ub : index, %step : index) {
+
+ // expected-error @below {{`omp.loop` expected to have a single nested operation which is a `omp.loop_nest`}}
+ omp.loop {
+ omp.simd {
+ omp.loop_nest (%iv) : index = (%lb) to (%ub) step (%step) {
+ omp.yield
+ }
+ } {omp.composite}
+ }
+
+ return
+}
+
+// -----
+
+func.func @omp_loop_invalid_nesting(%lb : index, %ub : index, %step : index) {
+
+ // expected-error @below {{custom op 'omp.loop' invalid clause value: 'dummy_value'}}
+ omp.loop bind(dummy_value) {
+ omp.loop_nest (%iv) : index = (%lb) to (%ub) step (%step) {
+ omp.yield
+ }
+ }
+
+ return
+}
diff --git a/mlir/test/Dialect/OpenMP/ops.mlir b/mlir/test/Dialect/OpenMP/ops.mlir
index 6f11b451fa00a3..47b8890b851c10 100644
--- a/mlir/test/Dialect/OpenMP/ops.mlir
+++ b/mlir/test/Dialect/OpenMP/ops.mlir
@@ -2749,3 +2749,43 @@ func.func @omp_target_private(%map1: memref<?xi32>, %map2: memref<?xi32>, %priv_
return
}
+
+// CHECK-LABEL: omp_loop
+func.func @omp_loop(%lb : index, %ub : index, %step : index) {
+ // CHECK: omp.loop {
+ omp.loop {
+ // CHECK: omp.loop_nest {{.*}} {
+ omp.loop_nest (%iv) : index = (%lb) to (%ub) step (%step) {
+ // CHECK: omp.yield
+ omp.yield
+ // CHECK: }
+ }
+ // CHECK: }
+ }
+
+ // CHECK: omp.loop bind(teams) {
+ omp.loop bind(teams) {
+ omp.loop_nest (%iv) : index = (%lb) to (%ub) step (%step) {
+ omp.yield
+ }
+ // CHECK: }
+ }
+
+ // CHECK: omp.loop bind(parallel) {
+ omp.loop bind(parallel) {
+ omp.loop_nest (%iv) : index = (%lb) to (%ub) step (%step) {
+ omp.yield
+ }
+ // CHECK: }
+ }
+
+ // CHECK: omp.loop bind(thread) {
+ omp.loop bind(thread) {
+ omp.loop_nest (%iv) : index = (%lb) to (%ub) step (%step) {
+ omp.yield
+ }
+ // CHECK: }
+ }
+
+ return
+}
More information about the Mlir-commits
mailing list