07ed8187ac
Specifying a kernel with the `ptx_kernel` or `amdgpu_kernel` calling convention is a more idiomatic and compile-time performant than using the `nvvm.annoation !"kernel"` metadata. Transition OMPIRBuilder to use calling conventions for PTX kernels and no longer emit `nvvm.annoation`. Update OpenMPOpt to work with kernels specified via calling convention as well as metadata. Update OpenMP tests to use the calling conventions.
27 lines
769 B
LLVM
27 lines
769 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 ptx_kernel void @openmp_kernel() "kernel" {
|
|
; CHECK-LABEL: define weak ptx_kernel void @openmp_kernel(
|
|
; CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
ret void
|
|
}
|
|
|
|
define weak_odr ptx_kernel void @non_openmp_kernel() {
|
|
; CHECK-LABEL: define weak_odr ptx_kernel void @non_openmp_kernel() {
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
ret void
|
|
}
|
|
|
|
!llvm.module.flags = !{!0, !1}
|
|
|
|
!0 = !{i32 7, !"openmp", i32 51}
|
|
!1 = !{i32 7, !"openmp-device", i32 51}
|