Skip to content

[flang][cuda] Allow AbstractResult to run in gpu.module #118529

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Dec 3, 2024
Merged

[flang][cuda] Allow AbstractResult to run in gpu.module #118529

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions flang/include/flang/Optimizer/Passes/Pipelines.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@
#include "flang/Tools/CrossToolHelpers.h"
#include "mlir/Conversion/ReconcileUnrealizedCasts/ReconcileUnrealizedCasts.h"
#include "mlir/Conversion/SCFToControlFlow/SCFToControlFlow.h"
#include "mlir/Dialect/GPU/IR/GPUDialect.h"
#include "mlir/Dialect/LLVMIR/LLVMAttrs.h"
#include "mlir/Pass/PassManager.h"
#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
Expand Down
3 changes: 2 additions & 1 deletion flang/lib/Optimizer/Passes/Pipelines.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,8 @@ namespace fir {
void addNestedPassToAllTopLevelOperations(mlir::PassManager &pm,
PassConstructor ctor) {
addNestedPassToOps<mlir::func::FuncOp, mlir::omp::DeclareReductionOp,
mlir::omp::PrivateClauseOp, fir::GlobalOp>(pm, ctor);
mlir::omp::PrivateClauseOp, fir::GlobalOp,
mlir::gpu::GPUModuleOp>(pm, ctor);
}

void addNestedPassToAllTopLevelOperationsConditionally(
Expand Down
28 changes: 25 additions & 3 deletions flang/lib/Optimizer/Transforms/AbstractResult.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@
#include "flang/Optimizer/Dialect/Support/FIRContext.h"
#include "flang/Optimizer/Transforms/Passes.h"
#include "mlir/Dialect/Func/IR/FuncOps.h"
#include "mlir/Dialect/GPU/IR/GPUDialect.h"
#include "mlir/IR/Diagnostics.h"
#include "mlir/Pass/Pass.h"
#include "mlir/Pass/PassManager.h"
Expand Down Expand Up @@ -331,9 +332,10 @@ class AbstractResultOpt
using fir::impl::AbstractResultOptBase<
AbstractResultOpt>::AbstractResultOptBase;

void runOnSpecificOperation(mlir::func::FuncOp func, bool shouldBoxResult,
mlir::RewritePatternSet &patterns,
mlir::ConversionTarget &target) {
template <typename OpTy>
void runOnFunctionLikeOperation(OpTy func, bool shouldBoxResult,
mlir::RewritePatternSet &patterns,
mlir::ConversionTarget &target) {
auto loc = func.getLoc();
auto *context = &getContext();
// Convert function type itself if it has an abstract result.
Expand Down Expand Up @@ -384,6 +386,18 @@ class AbstractResultOpt
}
}

void runOnSpecificOperation(mlir::func::FuncOp func, bool shouldBoxResult,
mlir::RewritePatternSet &patterns,
mlir::ConversionTarget &target) {
runOnFunctionLikeOperation(func, shouldBoxResult, patterns, target);
}

void runOnSpecificOperation(mlir::gpu::GPUFuncOp func, bool shouldBoxResult,
mlir::RewritePatternSet &patterns,
mlir::ConversionTarget &target) {
runOnFunctionLikeOperation(func, shouldBoxResult, patterns, target);
}

inline static bool containsFunctionTypeWithAbstractResult(mlir::Type type) {
return mlir::TypeSwitch<mlir::Type, bool>(type)
.Case([](fir::BoxProcType boxProc) {
Expand Down Expand Up @@ -448,6 +462,14 @@ class AbstractResultOpt
mlir::TypeSwitch<mlir::Operation *, void>(op)
.Case<mlir::func::FuncOp, fir::GlobalOp>([&](auto op) {
runOnSpecificOperation(op, shouldBoxResult, patterns, target);
})
.Case<mlir::gpu::GPUModuleOp>([&](auto op) {
auto gpuMod = mlir::dyn_cast<mlir::gpu::GPUModuleOp>(*op);
for (auto funcOp : gpuMod.template getOps<mlir::func::FuncOp>())
runOnSpecificOperation(funcOp, shouldBoxResult, patterns, target);
for (auto gpuFuncOp : gpuMod.template getOps<mlir::gpu::GPUFuncOp>())
runOnSpecificOperation(gpuFuncOp, shouldBoxResult, patterns,
target);
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is it possible to have globals inside the GPU module?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes we duplicate globals in the gpu module. I'll need to add that too.

});

// Convert the calls and, if needed, the ReturnOp in the function body.
Expand Down
9 changes: 7 additions & 2 deletions flang/test/Driver/bbc-mlir-pass-pipeline.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,12 +17,14 @@
! CHECK-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
! CHECK-NEXT: (S) 0 num-dce'd - Number of operations DCE'd

! CHECK-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! CHECK-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! CHECK-NEXT: 'fir.global' Pipeline
! CHECK-NEXT: CharacterConversion
! CHECK-NEXT: 'func.func' Pipeline
! CHECK-NEXT: ArrayValueCopy
! CHECK-NEXT: CharacterConversion
! CHECK-NEXT: 'gpu.module' Pipeline
! CHECK-NEXT: CharacterConversion
! CHECK-NEXT: 'omp.declare_reduction' Pipeline
! CHECK-NEXT: CharacterConversion
! CHECK-NEXT: 'omp.private' Pipeline
Expand All @@ -48,13 +50,16 @@
! CHECK-NEXT: PolymorphicOpConversion
! CHECK-NEXT: AssumedRankOpConversion

! CHECK-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! CHECK-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! CHECK-NEXT: 'fir.global' Pipeline
! CHECK-NEXT: StackReclaim
! CHECK-NEXT: CFGConversion
! CHECK-NEXT: 'func.func' Pipeline
! CHECK-NEXT: StackReclaim
! CHECK-NEXT: CFGConversion
! CHECK-NEXT: 'gpu.module' Pipeline
! CHECK-NEXT: StackReclaim
! CHECK-NEXT: CFGConversion
! CHECK-NEXT: 'omp.declare_reduction' Pipeline
! CHECK-NEXT: StackReclaim
! CHECK-NEXT: CFGConversion
Expand Down
17 changes: 13 additions & 4 deletions flang/test/Driver/mlir-debug-pass-pipeline.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,11 +28,13 @@
! ALL: Pass statistics report

! ALL: Fortran::lower::VerifierPass
! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: 'fir.global' Pipeline
! ALL-NEXT: InlineElementals
! ALL-NEXT: 'func.func' Pipeline
! ALL-NEXT: InlineElementals
! ALL-NEXT: 'gpu.module' Pipeline
! ALL-NEXT: InlineElementals
! ALL-NEXT: 'omp.declare_reduction' Pipeline
! ALL-NEXT: InlineElementals
! ALL-NEXT: 'omp.private' Pipeline
Expand All @@ -49,12 +51,14 @@
! ALL-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
! ALL-NEXT: (S) 0 num-dce'd - Number of operations DCE'd

! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: 'fir.global' Pipeline
! ALL-NEXT: CharacterConversion
! ALL-NEXT: 'func.func' Pipeline
! ALL-NEXT: ArrayValueCopy
! ALL-NEXT: CharacterConversion
! ALL-NEXT: 'gpu.module' Pipeline
! ALL-NEXT: CharacterConversion
! ALL-NEXT: 'omp.declare_reduction' Pipeline
! ALL-NEXT: CharacterConversion
! ALL-NEXT: 'omp.private' Pipeline
Expand All @@ -78,13 +82,16 @@
! ALL-NEXT: PolymorphicOpConversion
! ALL-NEXT: AssumedRankOpConversion

! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: 'fir.global' Pipeline
! ALL-NEXT: StackReclaim
! ALL-NEXT: CFGConversion
! ALL-NEXT: 'func.func' Pipeline
! ALL-NEXT: StackReclaim
! ALL-NEXT: CFGConversion
! ALL-NEXT: 'gpu.module' Pipeline
! ALL-NEXT: StackReclaim
! ALL-NEXT: CFGConversion
! ALL-NEXT: 'omp.declare_reduction' Pipeline
! ALL-NEXT: StackReclaim
! ALL-NEXT: CFGConversion
Expand All @@ -99,11 +106,13 @@
! ALL-NEXT: (S) 0 num-dce'd - Number of operations DCE'd
! ALL-NEXT: BoxedProcedurePass

! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: 'fir.global' Pipeline
! ALL-NEXT: AbstractResultOpt
! ALL-NEXT: 'func.func' Pipeline
! ALL-NEXT: AbstractResultOpt
! ALL-NEXT: 'gpu.module' Pipeline
! ALL-NEXT: AbstractResultOpt
! ALL-NEXT: 'omp.declare_reduction' Pipeline
! ALL-NEXT: AbstractResultOpt
! ALL-NEXT: 'omp.private' Pipeline
Expand Down
22 changes: 17 additions & 5 deletions flang/test/Driver/mlir-pass-pipeline.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,13 +16,16 @@

! ALL: Fortran::lower::VerifierPass
! O2-NEXT: Canonicalizer
! ALL: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! ALL: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT:'fir.global' Pipeline
! O2-NEXT: SimplifyHLFIRIntrinsics
! ALL: InlineElementals
! ALL-NEXT:'func.func' Pipeline
! O2-NEXT: SimplifyHLFIRIntrinsics
! ALL: InlineElementals
! ALL-NEXT:'gpu.module' Pipeline
! O2-NEXT: SimplifyHLFIRIntrinsics
! ALL: InlineElementals
! ALL-NEXT:'omp.declare_reduction' Pipeline
! O2-NEXT: SimplifyHLFIRIntrinsics
! ALL: InlineElementals
Expand All @@ -33,11 +36,13 @@
! O2-NEXT: CSE
! O2-NEXT: (S) {{.*}} num-cse'd
! O2-NEXT: (S) {{.*}} num-dce'd
! O2-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! O2-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! O2-NEXT: 'fir.global' Pipeline
! O2-NEXT: OptimizedBufferization
! O2-NEXT: 'func.func' Pipeline
! O2-NEXT: OptimizedBufferization
! O2-NEXT: 'gpu.module' Pipeline
! O2-NEXT: OptimizedBufferization
! O2-NEXT: 'omp.declare_reduction' Pipeline
! O2-NEXT: OptimizedBufferization
! O2-NEXT: 'omp.private' Pipeline
Expand All @@ -54,12 +59,14 @@
! ALL-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
! ALL-NEXT: (S) 0 num-dce'd - Number of operations DCE'd

! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: 'fir.global' Pipeline
! ALL-NEXT: CharacterConversion
! ALL-NEXT: 'func.func' Pipeline
! ALL-NEXT: ArrayValueCopy
! ALL-NEXT: CharacterConversion
! ALL-NEXT: 'gpu.module' Pipeline
! ALL-NEXT: CharacterConversion
! ALL-NEXT: 'omp.declare_reduction' Pipeline
! ALL-NEXT: CharacterConversion
! ALL-NEXT: 'omp.private' Pipeline
Expand All @@ -86,13 +93,16 @@
! ALL-NEXT: AssumedRankOpConversion
! O2-NEXT: AddAliasTags

! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: 'fir.global' Pipeline
! ALL-NEXT: StackReclaim
! ALL-NEXT: CFGConversion
! ALL-NEXT: 'func.func' Pipeline
! ALL-NEXT: StackReclaim
! ALL-NEXT: CFGConversion
! ALL-NEXT: 'gpu.module' Pipeline
! ALL-NEXT: StackReclaim
! ALL-NEXT: CFGConversion
! ALL-NEXT: 'omp.declare_reduction' Pipeline
! ALL-NEXT: StackReclaim
! ALL-NEXT: CFGConversion
Expand All @@ -108,11 +118,13 @@
! ALL-NEXT: (S) 0 num-dce'd - Number of operations DCE'd
! ALL-NEXT: BoxedProcedurePass

! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
! ALL-NEXT: 'fir.global' Pipeline
! ALL-NEXT: AbstractResultOpt
! ALL-NEXT: 'func.func' Pipeline
! ALL-NEXT: AbstractResultOpt
! ALL-NEXT: 'gpu.module' Pipeline
! ALL-NEXT: AbstractResultOpt
! ALL-NEXT: 'omp.declare_reduction' Pipeline
! ALL-NEXT: AbstractResultOpt
! ALL-NEXT: 'omp.private' Pipeline
Expand Down
22 changes: 17 additions & 5 deletions flang/test/Fir/basic-program.fir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,13 +17,16 @@ func.func @_QQmain() {
// PASSES: Pass statistics report

// PASSES: Canonicalizer
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: 'fir.global' Pipeline
// PASSES-NEXT: SimplifyHLFIRIntrinsics
// PASSES-NEXT: InlineElementals
// PASSES-NEXT: 'func.func' Pipeline
// PASSES-NEXT: SimplifyHLFIRIntrinsics
// PASSES-NEXT: InlineElementals
// PASSES-NEXT: 'gpu.module' Pipeline
// PASSES-NEXT: SimplifyHLFIRIntrinsics
// PASSES-NEXT: InlineElementals
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
// PASSES-NEXT: SimplifyHLFIRIntrinsics
// PASSES-NEXT: InlineElementals
Expand All @@ -34,11 +37,13 @@ func.func @_QQmain() {
// PASSES-NEXT: CSE
// PASSES-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
// PASSES-NEXT: (S) 0 num-dce'd - Number of operations DCE'd
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: 'fir.global' Pipeline
// PASSES-NEXT: OptimizedBufferization
// PASSES-NEXT: 'func.func' Pipeline
// PASSES-NEXT: OptimizedBufferization
// PASSES-NEXT: 'gpu.module' Pipeline
// PASSES-NEXT: OptimizedBufferization
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
// PASSES-NEXT: OptimizedBufferization
// PASSES-NEXT: 'omp.private' Pipeline
Expand All @@ -52,12 +57,14 @@ func.func @_QQmain() {
// PASSES-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
// PASSES-NEXT: (S) 0 num-dce'd - Number of operations DCE'd

// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: 'fir.global' Pipeline
// PASSES-NEXT: CharacterConversion
// PASSES-NEXT: 'func.func' Pipeline
// PASSES-NEXT: ArrayValueCopy
// PASSES-NEXT: CharacterConversion
// PASSES-NEXT: 'gpu.module' Pipeline
// PASSES-NEXT: CharacterConversion
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
// PASSES-NEXT: CharacterConversion
// PASSES-NEXT: 'omp.private' Pipeline
Expand All @@ -84,13 +91,16 @@ func.func @_QQmain() {
// PASSES-NEXT: AssumedRankOpConversion
// PASSES-NEXT: AddAliasTags

// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: 'fir.global' Pipeline
// PASSES-NEXT: StackReclaim
// PASSES-NEXT: CFGConversion
// PASSES-NEXT: 'func.func' Pipeline
// PASSES-NEXT: StackReclaim
// PASSES-NEXT: CFGConversion
// PASSES-NEXT: 'gpu.module' Pipeline
// PASSES-NEXT: StackReclaim
// PASSES-NEXT: CFGConversion
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
// PASSES-NEXT: StackReclaim
// PASSES-NEXT: CFGConversion
Expand All @@ -106,11 +116,13 @@ func.func @_QQmain() {
// PASSES-NEXT: (S) 0 num-dce'd - Number of operations DCE'd
// PASSES-NEXT: BoxedProcedurePass

// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
// PASSES-NEXT: 'fir.global' Pipeline
// PASSES-NEXT: AbstractResultOpt
// PASSES-NEXT: 'func.func' Pipeline
// PASSES-NEXT: AbstractResultOpt
// PASSES-NEXT: 'gpu.module' Pipeline
// PASSES-NEXT: AbstractResultOpt
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
// PASSES-NEXT: AbstractResultOpt
// PASSES-NEXT: 'omp.private' Pipeline
Expand Down
Loading