[OpenMP]Support for lowering masked op #98401
Merged
Conversation
This file contains 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
|
@llvm/pr-subscribers-mlir @llvm/pr-subscribers-flang-openmp Author: Anchu Rajendran S (anchuraj) ChangesChange adds the lowering support for Omp masked directive. Omp masked is introduced in 5.2 standard and allows a parallel region to be executed by threads specified by a programmer. This is achieved with the help of filter clause which helps to specify thread id expected to execute the region. Other previous related PRs: Full diff: https://github.com/llvm/llvm-project/pull/98401.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 0c9c699a1f390..591900723e2e9 100644
--- a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
+++ b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
@@ -264,6 +264,43 @@ static llvm::omp::ProcBindKind getProcBindKind(omp::ClauseProcBindKind kind) {
llvm_unreachable("Unknown ClauseProcBindKind kind");
}
+/// Converts an OpenMP 'masked' operation into LLVM IR using OpenMPIRBuilder.
+static LogicalResult
+convertOmpMasked(Operation &opInst, llvm::IRBuilderBase &builder,
+ LLVM::ModuleTranslation &moduleTranslation) {
+ auto maskedOp = cast<omp::MaskedOp>(opInst);
+ using InsertPointTy = llvm::OpenMPIRBuilder::InsertPointTy;
+ // TODO: support error propagation in OpenMPIRBuilder and use it instead of
+ // relying on captured variables.
+ LogicalResult bodyGenStatus = success();
+
+ auto bodyGenCB = [&](InsertPointTy allocaIP, InsertPointTy codeGenIP) {
+ // MaskedOp has only one region associated with it.
+ auto ®ion = maskedOp.getRegion();
+ builder.restoreIP(codeGenIP);
+ convertOmpOpRegions(region, "omp.masked.region", builder, moduleTranslation,
+ bodyGenStatus);
+ };
+
+ // TODO: Perform finalization actions for variables. This has to be
+ // called for variables which have destructors/finalizers.
+ auto finiCB = [&](InsertPointTy codeGenIP) {};
+
+ llvm::Value *filterVal = nullptr;
+ if (auto filterVar = maskedOp.getFilteredThreadId()) {
+ filterVal = moduleTranslation.lookupValue(filterVar);
+ } else {
+ llvm::LLVMContext &llvmContext = builder.getContext();
+ filterVal =
+ llvm::ConstantInt::get(llvm::Type::getInt32Ty(llvmContext), /*V=*/0);
+ }
+ assert(filterVal != nullptr);
+ llvm::OpenMPIRBuilder::LocationDescription ompLoc(builder);
+ builder.restoreIP(moduleTranslation.getOpenMPBuilder()->createMasked(
+ ompLoc, bodyGenCB, finiCB, filterVal));
+ return success();
+}
+
/// Converts an OpenMP 'master' operation into LLVM IR using OpenMPIRBuilder.
static LogicalResult
convertOmpMaster(Operation &opInst, llvm::IRBuilderBase &builder,
@@ -3330,6 +3367,9 @@ convertHostOrTargetOperation(Operation *op, llvm::IRBuilderBase &builder,
.Case([&](omp::ParallelOp op) {
return convertOmpParallel(op, builder, moduleTranslation);
})
+ .Case([&](omp::MaskedOp) {
+ return convertOmpMasked(*op, builder, moduleTranslation);
+ })
.Case([&](omp::MasterOp) {
return convertOmpMaster(*op, builder, moduleTranslation);
})
diff --git a/mlir/test/Target/LLVMIR/openmp-llvm.mlir b/mlir/test/Target/LLVMIR/openmp-llvm.mlir
index dfeaf4be33adb..188e570d01465 100644
--- a/mlir/test/Target/LLVMIR/openmp-llvm.mlir
+++ b/mlir/test/Target/LLVMIR/openmp-llvm.mlir
@@ -310,6 +310,35 @@ llvm.func @test_omp_master() -> () {
// -----
+// CHECK-LABEL: define void @test_omp_masked({{.*}})
+llvm.func @test_omp_masked(%arg0: i32)-> () {
+// CHECK: call void {{.*}}@__kmpc_fork_call{{.*}} @{{.*}})
+// CHECK: omp.par.region1:
+ omp.parallel {
+ omp.masked filter(%arg0: i32) {
+// CHECK: [[OMP_THREAD_3_4:%.*]] = call i32 @__kmpc_global_thread_num(ptr @{{[0-9]+}})
+// CHECK: {{[0-9]+}} = call i32 @__kmpc_masked(ptr @{{[0-9]+}}, i32 [[OMP_THREAD_3_4]], i32 %{{[0-9]+}})
+// CHECK: omp.masked.region
+// CHECK: call void @__kmpc_end_masked(ptr @{{[0-9]+}}, i32 [[OMP_THREAD_3_4]])
+// CHECK: br label %omp_region.end
+ omp.terminator
+ }
+ omp.terminator
+ }
+ omp.parallel {
+ omp.parallel {
+ omp.masked filter(%arg0: i32){
+ omp.terminator
+ }
+ omp.terminator
+ }
+ omp.terminator
+ }
+ llvm.return
+}
+
+// -----
+
// CHECK: %struct.ident_t = type
// CHECK: @[[$loc:.*]] = private unnamed_addr constant {{.*}} c";unknown;unknown;{{[0-9]+}};{{[0-9]+}};;\00"
// CHECK: @[[$loc_struct:.*]] = private unnamed_addr constant %struct.ident_t {{.*}} @[[$loc]] {{.*}}
|
|
@llvm/pr-subscribers-mlir-openmp Author: Anchu Rajendran S (anchuraj) ChangesChange adds the lowering support for Omp masked directive. Omp masked is introduced in 5.2 standard and allows a parallel region to be executed by threads specified by a programmer. This is achieved with the help of filter clause which helps to specify thread id expected to execute the region. Other previous related PRs: Full diff: https://github.com/llvm/llvm-project/pull/98401.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 0c9c699a1f390..591900723e2e9 100644
--- a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
+++ b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
@@ -264,6 +264,43 @@ static llvm::omp::ProcBindKind getProcBindKind(omp::ClauseProcBindKind kind) {
llvm_unreachable("Unknown ClauseProcBindKind kind");
}
+/// Converts an OpenMP 'masked' operation into LLVM IR using OpenMPIRBuilder.
+static LogicalResult
+convertOmpMasked(Operation &opInst, llvm::IRBuilderBase &builder,
+ LLVM::ModuleTranslation &moduleTranslation) {
+ auto maskedOp = cast<omp::MaskedOp>(opInst);
+ using InsertPointTy = llvm::OpenMPIRBuilder::InsertPointTy;
+ // TODO: support error propagation in OpenMPIRBuilder and use it instead of
+ // relying on captured variables.
+ LogicalResult bodyGenStatus = success();
+
+ auto bodyGenCB = [&](InsertPointTy allocaIP, InsertPointTy codeGenIP) {
+ // MaskedOp has only one region associated with it.
+ auto ®ion = maskedOp.getRegion();
+ builder.restoreIP(codeGenIP);
+ convertOmpOpRegions(region, "omp.masked.region", builder, moduleTranslation,
+ bodyGenStatus);
+ };
+
+ // TODO: Perform finalization actions for variables. This has to be
+ // called for variables which have destructors/finalizers.
+ auto finiCB = [&](InsertPointTy codeGenIP) {};
+
+ llvm::Value *filterVal = nullptr;
+ if (auto filterVar = maskedOp.getFilteredThreadId()) {
+ filterVal = moduleTranslation.lookupValue(filterVar);
+ } else {
+ llvm::LLVMContext &llvmContext = builder.getContext();
+ filterVal =
+ llvm::ConstantInt::get(llvm::Type::getInt32Ty(llvmContext), /*V=*/0);
+ }
+ assert(filterVal != nullptr);
+ llvm::OpenMPIRBuilder::LocationDescription ompLoc(builder);
+ builder.restoreIP(moduleTranslation.getOpenMPBuilder()->createMasked(
+ ompLoc, bodyGenCB, finiCB, filterVal));
+ return success();
+}
+
/// Converts an OpenMP 'master' operation into LLVM IR using OpenMPIRBuilder.
static LogicalResult
convertOmpMaster(Operation &opInst, llvm::IRBuilderBase &builder,
@@ -3330,6 +3367,9 @@ convertHostOrTargetOperation(Operation *op, llvm::IRBuilderBase &builder,
.Case([&](omp::ParallelOp op) {
return convertOmpParallel(op, builder, moduleTranslation);
})
+ .Case([&](omp::MaskedOp) {
+ return convertOmpMasked(*op, builder, moduleTranslation);
+ })
.Case([&](omp::MasterOp) {
return convertOmpMaster(*op, builder, moduleTranslation);
})
diff --git a/mlir/test/Target/LLVMIR/openmp-llvm.mlir b/mlir/test/Target/LLVMIR/openmp-llvm.mlir
index dfeaf4be33adb..188e570d01465 100644
--- a/mlir/test/Target/LLVMIR/openmp-llvm.mlir
+++ b/mlir/test/Target/LLVMIR/openmp-llvm.mlir
@@ -310,6 +310,35 @@ llvm.func @test_omp_master() -> () {
// -----
+// CHECK-LABEL: define void @test_omp_masked({{.*}})
+llvm.func @test_omp_masked(%arg0: i32)-> () {
+// CHECK: call void {{.*}}@__kmpc_fork_call{{.*}} @{{.*}})
+// CHECK: omp.par.region1:
+ omp.parallel {
+ omp.masked filter(%arg0: i32) {
+// CHECK: [[OMP_THREAD_3_4:%.*]] = call i32 @__kmpc_global_thread_num(ptr @{{[0-9]+}})
+// CHECK: {{[0-9]+}} = call i32 @__kmpc_masked(ptr @{{[0-9]+}}, i32 [[OMP_THREAD_3_4]], i32 %{{[0-9]+}})
+// CHECK: omp.masked.region
+// CHECK: call void @__kmpc_end_masked(ptr @{{[0-9]+}}, i32 [[OMP_THREAD_3_4]])
+// CHECK: br label %omp_region.end
+ omp.terminator
+ }
+ omp.terminator
+ }
+ omp.parallel {
+ omp.parallel {
+ omp.masked filter(%arg0: i32){
+ omp.terminator
+ }
+ omp.terminator
+ }
+ omp.terminator
+ }
+ llvm.return
+}
+
+// -----
+
// CHECK: %struct.ident_t = type
// CHECK: @[[$loc:.*]] = private unnamed_addr constant {{.*}} c";unknown;unknown;{{[0-9]+}};{{[0-9]+}};;\00"
// CHECK: @[[$loc_struct:.*]] = private unnamed_addr constant %struct.ident_t {{.*}} @[[$loc]] {{.*}}
|
|
@llvm/pr-subscribers-mlir-llvm Author: Anchu Rajendran S (anchuraj) ChangesChange adds the lowering support for Omp masked directive. Omp masked is introduced in 5.2 standard and allows a parallel region to be executed by threads specified by a programmer. This is achieved with the help of filter clause which helps to specify thread id expected to execute the region. Other previous related PRs: Full diff: https://github.com/llvm/llvm-project/pull/98401.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 0c9c699a1f390..591900723e2e9 100644
--- a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
+++ b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
@@ -264,6 +264,43 @@ static llvm::omp::ProcBindKind getProcBindKind(omp::ClauseProcBindKind kind) {
llvm_unreachable("Unknown ClauseProcBindKind kind");
}
+/// Converts an OpenMP 'masked' operation into LLVM IR using OpenMPIRBuilder.
+static LogicalResult
+convertOmpMasked(Operation &opInst, llvm::IRBuilderBase &builder,
+ LLVM::ModuleTranslation &moduleTranslation) {
+ auto maskedOp = cast<omp::MaskedOp>(opInst);
+ using InsertPointTy = llvm::OpenMPIRBuilder::InsertPointTy;
+ // TODO: support error propagation in OpenMPIRBuilder and use it instead of
+ // relying on captured variables.
+ LogicalResult bodyGenStatus = success();
+
+ auto bodyGenCB = [&](InsertPointTy allocaIP, InsertPointTy codeGenIP) {
+ // MaskedOp has only one region associated with it.
+ auto ®ion = maskedOp.getRegion();
+ builder.restoreIP(codeGenIP);
+ convertOmpOpRegions(region, "omp.masked.region", builder, moduleTranslation,
+ bodyGenStatus);
+ };
+
+ // TODO: Perform finalization actions for variables. This has to be
+ // called for variables which have destructors/finalizers.
+ auto finiCB = [&](InsertPointTy codeGenIP) {};
+
+ llvm::Value *filterVal = nullptr;
+ if (auto filterVar = maskedOp.getFilteredThreadId()) {
+ filterVal = moduleTranslation.lookupValue(filterVar);
+ } else {
+ llvm::LLVMContext &llvmContext = builder.getContext();
+ filterVal =
+ llvm::ConstantInt::get(llvm::Type::getInt32Ty(llvmContext), /*V=*/0);
+ }
+ assert(filterVal != nullptr);
+ llvm::OpenMPIRBuilder::LocationDescription ompLoc(builder);
+ builder.restoreIP(moduleTranslation.getOpenMPBuilder()->createMasked(
+ ompLoc, bodyGenCB, finiCB, filterVal));
+ return success();
+}
+
/// Converts an OpenMP 'master' operation into LLVM IR using OpenMPIRBuilder.
static LogicalResult
convertOmpMaster(Operation &opInst, llvm::IRBuilderBase &builder,
@@ -3330,6 +3367,9 @@ convertHostOrTargetOperation(Operation *op, llvm::IRBuilderBase &builder,
.Case([&](omp::ParallelOp op) {
return convertOmpParallel(op, builder, moduleTranslation);
})
+ .Case([&](omp::MaskedOp) {
+ return convertOmpMasked(*op, builder, moduleTranslation);
+ })
.Case([&](omp::MasterOp) {
return convertOmpMaster(*op, builder, moduleTranslation);
})
diff --git a/mlir/test/Target/LLVMIR/openmp-llvm.mlir b/mlir/test/Target/LLVMIR/openmp-llvm.mlir
index dfeaf4be33adb..188e570d01465 100644
--- a/mlir/test/Target/LLVMIR/openmp-llvm.mlir
+++ b/mlir/test/Target/LLVMIR/openmp-llvm.mlir
@@ -310,6 +310,35 @@ llvm.func @test_omp_master() -> () {
// -----
+// CHECK-LABEL: define void @test_omp_masked({{.*}})
+llvm.func @test_omp_masked(%arg0: i32)-> () {
+// CHECK: call void {{.*}}@__kmpc_fork_call{{.*}} @{{.*}})
+// CHECK: omp.par.region1:
+ omp.parallel {
+ omp.masked filter(%arg0: i32) {
+// CHECK: [[OMP_THREAD_3_4:%.*]] = call i32 @__kmpc_global_thread_num(ptr @{{[0-9]+}})
+// CHECK: {{[0-9]+}} = call i32 @__kmpc_masked(ptr @{{[0-9]+}}, i32 [[OMP_THREAD_3_4]], i32 %{{[0-9]+}})
+// CHECK: omp.masked.region
+// CHECK: call void @__kmpc_end_masked(ptr @{{[0-9]+}}, i32 [[OMP_THREAD_3_4]])
+// CHECK: br label %omp_region.end
+ omp.terminator
+ }
+ omp.terminator
+ }
+ omp.parallel {
+ omp.parallel {
+ omp.masked filter(%arg0: i32){
+ omp.terminator
+ }
+ omp.terminator
+ }
+ omp.terminator
+ }
+ llvm.return
+}
+
+// -----
+
// CHECK: %struct.ident_t = type
// CHECK: @[[$loc:.*]] = private unnamed_addr constant {{.*}} c";unknown;unknown;{{[0-9]+}};{{[0-9]+}};;\00"
// CHECK: @[[$loc_struct:.*]] = private unnamed_addr constant %struct.ident_t {{.*}} @[[$loc]] {{.*}}
|
skatrak
approved these changes
Jul 11, 2024
anchuraj
added a commit
that referenced
this pull request
Jul 12, 2024
PR adds changes to the flang frontend to create the `MaskedOp` when `masked` directive is used in the input program. Omp masked is introduced in 5.2 standard and allows a parallel region to be executed by threads specified by a programmer. This is achieved with the help of filter clause which helps to specify thread id expected to execute the region. Other related PRs: - [Fortran Parsing and Semantic Support](#91432) - Merged - [MLIR Support](https://github.com/llvm/llvm-project/pull/96022/files) - Merged - [Lowering Support](#98401) - Under Review
aaryanshukla
pushed a commit
to aaryanshukla/llvm-project
that referenced
this pull request
Jul 14, 2024
aaryanshukla
pushed a commit
to aaryanshukla/llvm-project
that referenced
this pull request
Jul 14, 2024
PR adds changes to the flang frontend to create the `MaskedOp` when `masked` directive is used in the input program. Omp masked is introduced in 5.2 standard and allows a parallel region to be executed by threads specified by a programmer. This is achieved with the help of filter clause which helps to specify thread id expected to execute the region. Other related PRs: - [Fortran Parsing and Semantic Support](llvm#91432) - Merged - [MLIR Support](https://github.com/llvm/llvm-project/pull/96022/files) - Merged - [Lowering Support](llvm#98401) - Under Review
negril
added a commit
to negril/llvm-project
that referenced
this pull request
Aug 25, 2024
Changed in version 3.12:
A backslash-character pair that is not a valid escape sequence now generates a
SyntaxWarning, instead of DeprecationWarning. For example, re.compile("\d+\.
\d+") now emits a SyntaxWarning ("\d" is an invalid escape sequence, use raw
strings for regular expression: re.compile(r"\d+\.\d+")). In a future Python
version, SyntaxError will eventually be raised, instead of SyntaxWarning.
(Contributed by Victor Stinner in llvmgh-98401.)
Closes: llvm#97815
See-also: https://docs.python.org/3/reference/lexical_analysis.html#escape-sequences
Signed-off-by: Paul Zander <[email protected]>
negril
added a commit
to negril/llvm-project
that referenced
this pull request
Aug 25, 2024
Changed in version 3.12:
A backslash-character pair that is not a valid escape sequence now generates a
SyntaxWarning, instead of DeprecationWarning. For example, re.compile("\d+\.
\d+") now emits a SyntaxWarning ("\d" is an invalid escape sequence, use raw
strings for regular expression: re.compile(r"\d+\.\d+")). In a future Python
version, SyntaxError will eventually be raised, instead of SyntaxWarning.
(Contributed by Victor Stinner in llvmgh-98401.)
Closes: llvm#97815
See-also: https://docs.python.org/3/reference/lexical_analysis.html#escape-sequences
Signed-off-by: Paul Zander <[email protected]>
negril
added a commit
to negril/llvm-project
that referenced
this pull request
Aug 25, 2024
Changed in version 3.12:
A backslash-character pair that is not a valid escape sequence now generates a
SyntaxWarning, instead of DeprecationWarning. For example, re.compile("\d+\.
\d+") now emits a SyntaxWarning ("\d" is an invalid escape sequence, use raw
strings for regular expression: re.compile(r"\d+\.\d+")). In a future Python
version, SyntaxError will eventually be raised, instead of SyntaxWarning.
(Contributed by Victor Stinner in llvmgh-98401.)
Closes: llvm#97815
See-also: https://docs.python.org/3/reference/lexical_analysis.html#escape-sequences
Signed-off-by: Paul Zander <[email protected]>
negril
added a commit
to negril/llvm-project
that referenced
this pull request
Aug 25, 2024
Changed in python version 3.12:
A backslash-character pair that is not a valid escape sequence now generates a
SyntaxWarning, instead of DeprecationWarning. For example, re.compile("\d+\.
\d+") now emits a SyntaxWarning ("\d" is an invalid escape sequence, use raw
strings for regular expression: re.compile(r"\d+\.\d+")). In a future Python
version, SyntaxError will eventually be raised, instead of SyntaxWarning.
(Contributed by Victor Stinner in llvmgh-98401.)
Closes: llvm#97815
See-also: https://docs.python.org/3/reference/lexical_analysis.html#escape-sequences
Signed-off-by: Paul Zander <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Change adds the lowering support for Omp masked directive. Omp masked is introduced in 5.2 standard and allows a parallel region to be executed by threads specified by a programmer. This is achieved with the help of filter clause which helps to specify thread id expected to execute the region.
Other related PRs: