Skip to content

[flang][cuda] Only copy global that have effective use #137890

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
Apr 29, 2025
Merged

[flang][cuda] Only copy global that have effective use #137890

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
9 changes: 9 additions & 0 deletions flang/lib/Optimizer/Transforms/CUFDeviceGlobal.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,15 @@ static void processAddrOfOp(fir::AddrOfOp addrOfOp,
mlir::SymbolTable &symbolTable,
llvm::DenseSet<fir::GlobalOp> &candidates,
bool recurseInGlobal) {

// Check if there is a real use of the global.
if (addrOfOp.getOperation()->hasOneUse()) {
mlir::OpOperand &addrUse = *addrOfOp.getOperation()->getUses().begin();
if (mlir::isa<fir::DeclareOp>(addrUse.getOwner()) &&
addrUse.getOwner()->use_empty())
return;
}

if (auto globalOp = symbolTable.lookup<fir::GlobalOp>(
addrOfOp.getSymbol().getRootReference().getValue())) {
// TO DO: limit candidates to non-scalars. Scalars appear to have been
Expand Down
13 changes: 13 additions & 0 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 @@ -329,3 +329,16 @@ // attributes(global) subroutine kernel4()
// CHECK-LABEL: fir.global internal @_QFkernel4Ea : i32
// CHECK-LABEL: gpu.module @cuda_device_mod
// CHECK: fir.global internal @_QFkernel4Ea : i32

// -----

fir.global @_QMiso_c_bindingECc_alert constant : !fir.char<1>
func.func @_QMcudafor_lib_internalsPfoo() attributes {cuf.proc_attr = #cuf.cuda_proc<global>} {
%19 = fir.address_of(@_QMiso_c_bindingECc_alert) : !fir.ref<!fir.char<1>>
%c1 = arith.constant 1 : index
%20 = fir.declare %19 typeparams %c1 {fortran_attrs = #fir.var_attrs<parameter>, uniq_name = "_QMiso_c_bindingECc_alert"} : (!fir.ref<!fir.char<1>>, index) -> !fir.ref<!fir.char<1>>
return
}

// CHECK-LABEL: gpu.module @cuda_device_mod
// CHECK-NOT: _QMiso_c_bindingECc_alert