Skip to content

[flang][cuda] Bring PARAMETER arrays into the GPU module #146416

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
Jun 30, 2025
Merged

[flang][cuda] Bring PARAMETER arrays into the GPU module #146416

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
10 changes: 9 additions & 1 deletion flang/lib/Optimizer/Transforms/CUFDeviceGlobal.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -113,8 +113,16 @@ class CUFDeviceGlobal : public fir::impl::CUFDeviceGlobalBase<CUFDeviceGlobal> {
return signalPassFailure();
mlir::SymbolTable gpuSymTable(gpuMod);
for (auto globalOp : mod.getOps<fir::GlobalOp>()) {
if (cuf::isRegisteredDeviceGlobal(globalOp))
if (cuf::isRegisteredDeviceGlobal(globalOp)) {
candidates.insert(globalOp);
} else if (globalOp.getConstant() &&
mlir::isa<fir::SequenceType>(
fir::unwrapRefType(globalOp.resultType()))) {
mlir::Attribute initAttr =
globalOp.getInitVal().value_or(mlir::Attribute());
if (initAttr && mlir::dyn_cast<mlir::DenseElementsAttr>(initAttr))
candidates.insert(globalOp);
}
}
for (auto globalOp : candidates) {
auto globalName{globalOp.getSymbol().getValue()};
Expand Down
13 changes: 13 additions & 0 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 @@ -11,3 +11,16 @@ module attributes {fir.defaultkind = "a1c4d8i4l4r4", fir.kindmap = "", gpu.conta

// 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>

// -----

module attributes {fir.defaultkind = "a1c4d8i4l4r4", fir.kindmap = "", gpu.container_module} {
fir.global @_QMm1ECb(dense<[90, 100, 110]> : tensor<3xi32>) constant : !fir.array<3xi32>
fir.global @_QMm2ECc(dense<[100, 200, 300]> : tensor<3xi32>) constant : !fir.array<3xi32>
}

// CHECK: fir.global @_QMm1ECb
// CHECK: fir.global @_QMm2ECc
// CHECK: gpu.module @cuda_device_mod
// CHECK-DAG: fir.global @_QMm2ECc
// CHECK-DAG: fir.global @_QMm1ECb
Loading