[OpenMP][mlir] Added num_teams, thread_limit translation to LLVM IR
#68821
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This patch adds translation to LLVM IR for `num_teams` and `thread_limit` in for `omp.teams` operation.
shraiysh
requested review from
DominikAdamski,
NimishMishra,
TIFitis,
jsjodin,
kiranchandramohan,
kiranktp and
raghavendhra
|
@llvm/pr-subscribers-mlir @llvm/pr-subscribers-mlir-llvm Author: Shraiysh (shraiysh) ChangesThis patch adds translation to LLVM IR for Full diff: https://github.com/llvm/llvm-project/pull/68821.diff 2 Files Affected:
diff --git a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
index 1ec3bb8e7562a9e..ae974c14fac41a6 100644
--- a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
+++ b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
@@ -667,11 +667,9 @@ convertOmpTeams(omp::TeamsOp op, llvm::IRBuilderBase &builder,
LLVM::ModuleTranslation &moduleTranslation) {
using InsertPointTy = llvm::OpenMPIRBuilder::InsertPointTy;
LogicalResult bodyGenStatus = success();
- if (op.getNumTeamsLower() || op.getNumTeamsUpper() || op.getIfExpr() ||
- op.getThreadLimit() || !op.getAllocatorsVars().empty() ||
- op.getReductions()) {
+ if (op.getIfExpr() || !op.getAllocatorsVars().empty() || op.getReductions())
return op.emitError("unhandled clauses for translation to LLVM IR");
- }
+
auto bodyCB = [&](InsertPointTy allocaIP, InsertPointTy codegenIP) {
LLVM::ModuleTranslation::SaveStack<OpenMPAllocaStackFrame> frame(
moduleTranslation, allocaIP);
@@ -680,9 +678,21 @@ convertOmpTeams(omp::TeamsOp op, llvm::IRBuilderBase &builder,
moduleTranslation, bodyGenStatus);
};
+ llvm::Value *numTeamsLower = nullptr;
+ if (auto numTeamsLowerVar = op.getNumTeamsLower())
+ numTeamsLower = moduleTranslation.lookupValue(numTeamsLowerVar);
+
+ llvm::Value *numTeamsUpper = nullptr;
+ if (auto numTeamsUpperVar = op.getNumTeamsUpper())
+ numTeamsUpper = moduleTranslation.lookupValue(numTeamsUpperVar);
+
+ llvm::Value *threadLimit = nullptr;
+ if (auto threadLimitVar = op.getThreadLimit())
+ threadLimit = moduleTranslation.lookupValue(threadLimitVar);
+
llvm::OpenMPIRBuilder::LocationDescription ompLoc(builder);
- builder.restoreIP(
- moduleTranslation.getOpenMPBuilder()->createTeams(ompLoc, bodyCB));
+ builder.restoreIP(moduleTranslation.getOpenMPBuilder()->createTeams(
+ ompLoc, bodyCB, numTeamsLower, numTeamsUpper, threadLimit));
return bodyGenStatus;
}
diff --git a/mlir/test/Target/LLVMIR/openmp-teams.mlir b/mlir/test/Target/LLVMIR/openmp-teams.mlir
index 18fc2bb5a3c61b2..87ef90223ed704a 100644
--- a/mlir/test/Target/LLVMIR/openmp-teams.mlir
+++ b/mlir/test/Target/LLVMIR/openmp-teams.mlir
@@ -124,3 +124,114 @@ llvm.func @omp_teams_branching_shared(%condition: i1, %arg0: i32, %arg1: f32, %a
// CHECK-NEXT: br label
// CHECK: ret void
+// -----
+
+llvm.func @beforeTeams()
+llvm.func @duringTeams()
+llvm.func @afterTeams()
+
+// CHECK-LABEL: @omp_teams_thread_limit
+// CHECK-SAME: (i32 [[THREAD_LIMIT:.+]])
+llvm.func @omp_teams_thread_limit(%threadLimit: i32) {
+ // CHECK-NEXT: call void @beforeTeams()
+ llvm.call @beforeTeams() : () -> ()
+ // CHECK: [[THREAD_NUM:%.+]] = call i32 @__kmpc_global_thread_num
+ // CHECK-NEXT: call void @__kmpc_push_num_teams_51({{.+}}, i32 [[THREAD_NUM]], i32 0, i32 0, i32 [[THREAD_LIMIT]])
+ // CHECK: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @1, i32 0, ptr [[OUTLINED_FN:.+]])
+ omp.teams thread_limit(%threadLimit : i32) {
+ llvm.call @duringTeams() : () -> ()
+ omp.terminator
+ }
+ // CHECK: call void @afterTeams
+ llvm.call @afterTeams() : () -> ()
+ // CHECK: ret void
+ llvm.return
+}
+
+// CHECK: define internal void [[OUTLINED_FN]](ptr {{.+}}, ptr {{.+}})
+// CHECK: call void @duringTeams()
+// CHECK: ret void
+
+// -----
+
+llvm.func @beforeTeams()
+llvm.func @duringTeams()
+llvm.func @afterTeams()
+
+// CHECK-LABEL: @omp_teams_num_teams_upper
+// CHECK-SAME: (i32 [[NUM_TEAMS_UPPER:.+]])
+llvm.func @omp_teams_num_teams_upper(%numTeamsUpper: i32) {
+ // CHECK-NEXT: call void @beforeTeams()
+ llvm.call @beforeTeams() : () -> ()
+ // CHECK: [[THREAD_NUM:%.+]] = call i32 @__kmpc_global_thread_num
+ // CHECK-NEXT: call void @__kmpc_push_num_teams_51({{.+}}, i32 [[THREAD_NUM]], i32 [[NUM_TEAMS_UPPER]], i32 [[NUM_TEAMS_UPPER]], i32 0)
+ // CHECK: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @1, i32 0, ptr [[OUTLINED_FN:.+]])
+ omp.teams num_teams(to %numTeamsUpper : i32) {
+ llvm.call @duringTeams() : () -> ()
+ omp.terminator
+ }
+ // CHECK: call void @afterTeams
+ llvm.call @afterTeams() : () -> ()
+ // CHECK: ret void
+ llvm.return
+}
+
+// CHECK: define internal void [[OUTLINED_FN]](ptr {{.+}}, ptr {{.+}})
+// CHECK: call void @duringTeams()
+// CHECK: ret void
+
+// -----
+
+llvm.func @beforeTeams()
+llvm.func @duringTeams()
+llvm.func @afterTeams()
+
+// CHECK-LABEL: @omp_teams_num_teams_lower_and_upper
+// CHECK-SAME: (i32 [[NUM_TEAMS_LOWER:.+]], i32 [[NUM_TEAMS_UPPER:.+]])
+llvm.func @omp_teams_num_teams_lower_and_upper(%numTeamsLower: i32, %numTeamsUpper: i32) {
+ // CHECK-NEXT: call void @beforeTeams()
+ llvm.call @beforeTeams() : () -> ()
+ // CHECK: [[THREAD_NUM:%.+]] = call i32 @__kmpc_global_thread_num
+ // CHECK-NEXT: call void @__kmpc_push_num_teams_51({{.+}}, i32 [[THREAD_NUM]], i32 [[NUM_TEAMS_LOWER]], i32 [[NUM_TEAMS_UPPER]], i32 0)
+ // CHECK: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @1, i32 0, ptr [[OUTLINED_FN:.+]])
+ omp.teams num_teams(%numTeamsLower : i32 to %numTeamsUpper: i32) {
+ llvm.call @duringTeams() : () -> ()
+ omp.terminator
+ }
+ // CHECK: call void @afterTeams
+ llvm.call @afterTeams() : () -> ()
+ // CHECK: ret void
+ llvm.return
+}
+
+// CHECK: define internal void [[OUTLINED_FN]](ptr {{.+}}, ptr {{.+}})
+// CHECK: call void @duringTeams()
+// CHECK: ret void
+
+// -----
+
+llvm.func @beforeTeams()
+llvm.func @duringTeams()
+llvm.func @afterTeams()
+
+// CHECK-LABEL: @omp_teams_num_teams_and_thread_limit
+// CHECK-SAME: (i32 [[NUM_TEAMS_LOWER:.+]], i32 [[NUM_TEAMS_UPPER:.+]], i32 [[THREAD_LIMIT:.+]])
+llvm.func @omp_teams_num_teams_and_thread_limit(%numTeamsLower: i32, %numTeamsUpper: i32, %threadLimit: i32) {
+ // CHECK-NEXT: call void @beforeTeams()
+ llvm.call @beforeTeams() : () -> ()
+ // CHECK: [[THREAD_NUM:%.+]] = call i32 @__kmpc_global_thread_num
+ // CHECK-NEXT: call void @__kmpc_push_num_teams_51({{.+}}, i32 [[THREAD_NUM]], i32 [[NUM_TEAMS_LOWER]], i32 [[NUM_TEAMS_UPPER]], i32 [[THREAD_LIMIT]])
+ // CHECK: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @1, i32 0, ptr [[OUTLINED_FN:.+]])
+ omp.teams num_teams(%numTeamsLower : i32 to %numTeamsUpper: i32) thread_limit(%threadLimit: i32) {
+ llvm.call @duringTeams() : () -> ()
+ omp.terminator
+ }
+ // CHECK: call void @afterTeams
+ llvm.call @afterTeams() : () -> ()
+ // CHECK: ret void
+ llvm.return
+}
+
+// CHECK: define internal void [[OUTLINED_FN]](ptr {{.+}}, ptr {{.+}})
+// CHECK: call void @duringTeams()
+// CHECK: ret void
|
kiranchandramohan
approved these changes
Oct 12, 2023
Comment on lines
681
to
691
| llvm::Value *numTeamsLower = nullptr; | ||
| if (auto numTeamsLowerVar = op.getNumTeamsLower()) | ||
| numTeamsLower = moduleTranslation.lookupValue(numTeamsLowerVar); | ||
|
|
||
| llvm::Value *numTeamsUpper = nullptr; | ||
| if (auto numTeamsUpperVar = op.getNumTeamsUpper()) | ||
| numTeamsUpper = moduleTranslation.lookupValue(numTeamsUpperVar); | ||
|
|
||
| llvm::Value *threadLimit = nullptr; | ||
| if (auto threadLimitVar = op.getThreadLimit()) | ||
| threadLimit = moduleTranslation.lookupValue(threadLimitVar); |
There was a problem hiding this comment.
Nit: Spell the auto?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This patch adds translation to LLVM IR for
num_teamsandthread_limitin foromp.teamsoperation.