186a4b3b65
Current OpenMPOpt assumes all kernels are OpenMP kernels (aka. with "kernel" attribute). This doesn't hold if we mix OpenMP code and CUDA code by lingking them together because CUDA kernels are not annotated with the attribute. This patch removes the assumption and added a new counter for those non-OpenMP kernels. Fix #66687.
30 lines
849 B
LLVM
30 lines
849 B
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-attributes --version 3
|
|
; RUN: opt < %s -S -passes=openmp-opt | FileCheck %s
|
|
|
|
source_filename = "bug66687.ll"
|
|
target datalayout = "e-i64:64-i128:128-v16:16-v32:32-n16:32:64"
|
|
target triple = "nvptx64-nvidia-cuda"
|
|
|
|
define weak void @openmp_kernel() "kernel" {
|
|
; CHECK-LABEL: define weak void @openmp_kernel(
|
|
; CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
ret void
|
|
}
|
|
|
|
define weak_odr void @non_openmp_kernel() {
|
|
; CHECK-LABEL: define weak_odr void @non_openmp_kernel() {
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
ret void
|
|
}
|
|
|
|
!llvm.module.flags = !{!0, !1}
|
|
!nvvm.annotations = !{!2, !3}
|
|
|
|
!0 = !{i32 7, !"openmp", i32 51}
|
|
!1 = !{i32 7, !"openmp-device", i32 51}
|
|
!2 = !{ptr @openmp_kernel, !"kernel", i32 1}
|
|
!3 = !{ptr @non_openmp_kernel, !"kernel", i32 1}
|