Skip to content

[flang][cuda] Do not generate NVVM target attribute when creating the module #116882

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
Nov 20, 2024
Merged

[flang][cuda] Do not generate NVVM target attribute when creating the module #116882

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
3 changes: 0 additions & 3 deletions flang/lib/Optimizer/Transforms/CUFCommon.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,9 +22,6 @@ mlir::gpu::GPUModuleOp cuf::getOrCreateGPUModule(mlir::ModuleOp mod,
mlir::OpBuilder builder(ctx);
auto gpuMod = builder.create<mlir::gpu::GPUModuleOp>(mod.getLoc(),
cudaDeviceModuleName);
llvm::SmallVector<mlir::Attribute> targets;
targets.push_back(mlir::NVVM::NVVMTargetAttr::get(ctx));
gpuMod.setTargetsAttr(builder.getArrayAttr(targets));
mlir::Block::iterator insertPt(mod.getBodyRegion().front().end());
symTab.insert(gpuMod, insertPt);
return gpuMod;
Expand Down
2 changes: 1 addition & 1 deletion flang/test/Fir/CUDA/cuda-alloc-free.fir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,7 +73,7 @@ func.func @_QPtest_type() {
// CHECK: %[[CONV_BYTES:.*]] = fir.convert %[[BYTES]] : (index) -> i64
// CHECK: fir.call @_FortranACUFMemAlloc(%[[CONV_BYTES]], %c0{{.*}}, %{{.*}}, %{{.*}}) : (i64, i32, !fir.ref<i8>, i32) -> !fir.llvm_ptr<i8>

gpu.module @cuda_device_mod [#nvvm.target] {
gpu.module @cuda_device_mod {
gpu.func @_QMalloc() kernel {
%0 = cuf.alloc !fir.box<!fir.heap<!fir.array<?xf32>>> {bindc_name = "a", data_attr = #cuf.cuda<device>, uniq_name = "_QMallocEa"} -> !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
gpu.return
Expand Down
4 changes: 2 additions & 2 deletions flang/test/Fir/CUDA/cuda-constructor-2.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,11 +10,11 @@ module attributes {dlti.dl_spec = #dlti.dl_spec<#dlti.dl_entry<!llvm.ptr, dense<
fir.has_value %1 : !fir.box<!fir.heap<!fir.array<?xi32>>>
}

gpu.module @cuda_device_mod [#nvvm.target] {
gpu.module @cuda_device_mod {
}
}

// CHECK: gpu.module @cuda_device_mod [#nvvm.target]
// CHECK: gpu.module @cuda_device_mod

// CHECK: llvm.func internal @__cudaFortranConstructor() {
// CHECK-DAG: %[[MODULE:.*]] = cuf.register_module @cuda_device_mod -> !llvm.ptr
Expand Down
4 changes: 2 additions & 2 deletions flang/test/Fir/CUDA/cuda-device-global.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,9 +5,9 @@
module attributes {fir.defaultkind = "a1c4d8i4l4r4", fir.kindmap = "", gpu.container_module} {
fir.global @_QMmtestsEn(dense<[3, 4, 5, 6, 7]> : tensor<5xi32>) {data_attr = #cuf.cuda<device>} : !fir.array<5xi32>

gpu.module @cuda_device_mod [#nvvm.target] {
gpu.module @cuda_device_mod {
}
}

// CHECK: gpu.module @cuda_device_mod [#nvvm.target]
// CHECK: gpu.module @cuda_device_mo
// CHECK-NEXT: fir.global @_QMmtestsEn(dense<[3, 4, 5, 6, 7]> : tensor<5xi32>) {data_attr = #cuf.cuda<device>} : !fir.array<5xi32>
4 changes: 2 additions & 2 deletions flang/test/Fir/CUDA/cuda-implicit-device-global.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,7 @@ // Test that global used in device function are flagged with the correct
// CHECK: fir.call @_FortranAioBeginExternalListOutput(%{{.*}}, %[[CONV]], %{{.*}}) fastmath<contract> : (i32, !fir.ref<i8>, i32) -> !fir.ref<i8>
// CHECK: fir.global linkonce @_QQcl[[SYMBOL]] {data_attr = #cuf.cuda<constant>} constant : !fir.char<1,32>

// CHECK-LABEL: gpu.module @cuda_device_mod [#nvvm.target]
// CHECK-LABEL: gpu.module @cuda_device_mod
// CHECK: fir.global linkonce @_QQclX6995815537abaf90e86ce166af128f3a

// -----
Expand All @@ -51,5 +51,5 @@ // Test that global used in device function are flagged with the correct
// CHECK: fir.call @_FortranAioBeginExternalListOutput(%{{.*}}, %[[CONV]], %{{.*}}) fastmath<contract> : (i32, !fir.ref<i8>, i32) -> !fir.ref<i8>
// CHECK: fir.global linkonce @_QQcl[[SYMBOL]] constant : !fir.char<1,32>

// CHECK-LABEL: gpu.module @cuda_device_mod [#nvvm.target]
// CHECK-LABEL: gpu.module @cuda_device_mod
// CHECK-NOT: fir.global linkonce @_QQclX6995815537abaf90e86ce166af128f3a
Loading