Skip to content

[mlir][NVVM] Make sure barrier reduction attr can roundtrip #167958

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 13, 2025
+26 −19
Merged

[mlir][NVVM] Make sure barrier reduction attr can roundtrip #167958

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
18 changes: 9 additions & 9 deletions flang/test/Lower/CUDA/cuda-device-proc.cuf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -103,24 +103,24 @@ end
! CHECK-LABEL: func.func @_QPdevsub() attributes {cuf.proc_attr = #cuf.cuda_proc<global>}
! CHECK: nvvm.barrier0
! CHECK: nvvm.bar.warp.sync %c1{{.*}} : i32
! CHECK: %{{.*}} = nvvm.barrier <and> %c1{{.*}} -> i32
! CHECK: %{{.*}} = nvvm.barrier #nvvm.reduction<and> %c1{{.*}} -> i32
! CHECK: %[[A:.*]] = fir.load %{{.*}} : !fir.ref<i32>
! CHECK: %[[B:.*]] = fir.load %{{.*}} : !fir.ref<i32>
! CHECK: %[[CMP:.*]] = arith.cmpi sgt, %[[A]], %[[B]] : i32
! CHECK: %[[CONV:.*]] = fir.convert %[[CMP]] : (i1) -> i32
! CHECK: %{{.*}} = nvvm.barrier <and> %[[CONV]] -> i32
! CHECK: %{{.*}} = nvvm.barrier <popc> %c1{{.*}} -> i32
! CHECK: %{{.*}} = nvvm.barrier #nvvm.reduction<and> %[[CONV]] -> i32
! CHECK: %{{.*}} = nvvm.barrier #nvvm.reduction<popc> %c1{{.*}} -> i32
! CHECK: %[[A:.*]] = fir.load %{{.*}} : !fir.ref<i32>
! CHECK: %[[B:.*]] = fir.load %{{.*}} : !fir.ref<i32>
! CHECK: %[[CMP:.*]] = arith.cmpi sgt, %[[A]], %[[B]] : i32
! CHECK: %[[CONV:.*]] = fir.convert %[[CMP]] : (i1) -> i32
! CHECK: %{{.*}} = nvvm.barrier <popc> %[[CONV]] -> i32
! CHECK: %{{.*}} = nvvm.barrier <or> %c1{{.*}} -> i32
! CHECK: %{{.*}} = nvvm.barrier #nvvm.reduction<popc> %[[CONV]] -> i32
! CHECK: %{{.*}} = nvvm.barrier #nvvm.reduction<or> %c1{{.*}} -> i32
! CHECK: %[[A:.*]] = fir.load %{{.*}} : !fir.ref<i32>
! CHECK: %[[B:.*]] = fir.load %{{.*}} : !fir.ref<i32>
! CHECK: %[[CMP:.*]] = arith.cmpi sgt, %[[A]], %[[B]] : i32
! CHECK: %[[CONV:.*]] = fir.convert %[[CMP]] : (i1) -> i32
! CHECK: %{{.*}} = nvvm.barrier <or> %[[CONV]] -> i32
! CHECK: %{{.*}} = nvvm.barrier #nvvm.reduction<or> %[[CONV]] -> i32
! CHECK: %{{.*}} = llvm.atomicrmw add %{{.*}}, %{{.*}} seq_cst : !llvm.ptr, i32
! CHECK: %{{.*}} = llvm.atomicrmw add %{{.*}}, %{{.*}} seq_cst : !llvm.ptr, i64
! CHECK: %{{.*}} = llvm.atomicrmw fadd %{{.*}}, %{{.*}} seq_cst : !llvm.ptr, f32
Expand Down Expand Up @@ -214,9 +214,9 @@ end
! CHECK: cuf.kernel
! CHECK: nvvm.barrier0
! CHECK: nvvm.bar.warp.sync %c1{{.*}} : i32
! CHECK: nvvm.barrier <and> %c1{{.*}} -> i32
! CHECK: nvvm.barrier <popc> %c1{{.*}} -> i32
! CHECK: nvvm.barrier <or> %c1{{.*}} -> i32
! CHECK: nvvm.barrier #nvvm.reduction<and> %c1{{.*}} -> i32
! CHECK: nvvm.barrier #nvvm.reduction<popc> %c1{{.*}} -> i32
! CHECK: nvvm.barrier #nvvm.reduction<or> %c1{{.*}} -> i32

attributes(device) subroutine testMatch()
integer :: a, ipred, mask, v32
Expand Down
2 changes: 1 addition & 1 deletion mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -994,7 +994,7 @@ def NVVM_BarrierOp : NVVM_Op<"barrier", [AttrSizedOperandSegments]> {

let assemblyFormat =
"(`id` `=` $barrierId^)? (`number_of_threads` `=` $numberOfThreads^)? "
"($reductionOp^ $reductionPredicate)? (`->` type($res)^)? attr-dict";
"(qualified($reductionOp)^ $reductionPredicate)? (`->` type($res)^)? attr-dict";

let builders = [OpBuilder<(ins), [{
return build($_builder, $_state, TypeRange{}, Value{}, Value{}, {}, Value{});
Expand Down
25 changes: 16 additions & 9 deletions mlir/test/Target/LLVMIR/nvvm/barrier.mlir
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,19 +1,26 @@
// RUN: mlir-translate -mlir-to-llvmir %s -split-input-file --verify-diagnostics | FileCheck %s
// RUN: mlir-translate -mlir-to-llvmir %s -split-input-file --verify-diagnostics | FileCheck %s --check-prefix=LLVM
// RUN: mlir-opt %s -split-input-file | mlir-opt | FileCheck %s
Copy link
Contributor

@durga4github durga4github Nov 14, 2025

Choose a reason for hiding this comment

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

ok, we usually use the split-input-file for files with many negative tests (with a -- line separating each one). So, for this positive-test file, I think split-input-file and verify-diag do not seem very relevant. Could you please check once if removing them is all good, and if so, remove them?

Copy link
Contributor

@durga4github durga4github Nov 14, 2025

Choose a reason for hiding this comment

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

ok, nevermind, did not notice it was already merged.
We can clean this up when we are around here later


// CHECK-LABEL: @llvm_nvvm_barrier(
// CHECK-SAME: i32 %[[barId:.*]], i32 %[[numThreads:.*]], i32 %[[redOperand:.*]])
// LLVM-LABEL: @llvm_nvvm_barrier(
// LLVM-SAME: i32 %[[barId:.*]], i32 %[[numThreads:.*]], i32 %[[redOperand:.*]])
llvm.func @llvm_nvvm_barrier(%barID : i32, %numberOfThreads : i32, %redOperand : i32) {
// CHECK: call void @llvm.nvvm.barrier.cta.sync.aligned.all(i32 0)
// LLVM: call void @llvm.nvvm.barrier.cta.sync.aligned.all(i32 0)
// CHECK: nvvm.barrier
nvvm.barrier
// CHECK: call void @llvm.nvvm.barrier.cta.sync.aligned.all(i32 %[[barId]])
// LLVM: call void @llvm.nvvm.barrier.cta.sync.aligned.all(i32 %[[barId]])
// CHECK: nvvm.barrier id = %{{.*}}
nvvm.barrier id = %barID
// CHECK: call void @llvm.nvvm.barrier.cta.sync.aligned.count(i32 %[[barId]], i32 %[[numThreads]])
// LLVM: call void @llvm.nvvm.barrier.cta.sync.aligned.count(i32 %[[barId]], i32 %[[numThreads]])
// CHECK: nvvm.barrier id = %{{.*}} number_of_threads = %{{.*}}
nvvm.barrier id = %barID number_of_threads = %numberOfThreads
// CHECK: %{{.*}} = call i32 @llvm.nvvm.barrier0.and(i32 %[[redOperand]])
// LLVM: %{{.*}} = call i32 @llvm.nvvm.barrier0.and(i32 %[[redOperand]])
// CHECK: %{{.*}} = nvvm.barrier #nvvm.reduction<and> %{{.*}} -> i32
%0 = nvvm.barrier #nvvm.reduction<and> %redOperand -> i32
// CHECK: %{{.*}} = call i32 @llvm.nvvm.barrier0.or(i32 %[[redOperand]])
// LLVM: %{{.*}} = call i32 @llvm.nvvm.barrier0.or(i32 %[[redOperand]])
// CHECK: %{{.*}} = nvvm.barrier #nvvm.reduction<or> %{{.*}} -> i32
%1 = nvvm.barrier #nvvm.reduction<or> %redOperand -> i32
// CHECK: %{{.*}} = call i32 @llvm.nvvm.barrier0.popc(i32 %[[redOperand]])
// LLVM: %{{.*}} = call i32 @llvm.nvvm.barrier0.popc(i32 %[[redOperand]])
// CHECK: %{{.*}} = nvvm.barrier #nvvm.reduction<popc> %{{.*}} -> i32
%2 = nvvm.barrier #nvvm.reduction<popc> %redOperand -> i32

llvm.return
Expand Down
Loading