fc0653f31c
Recent upstream trends have moved away from explicitly using `-verify-machineinstrs`, as it's already covered by the expensive checks. This PR removes almost all `-verify-machineinstrs` from tests in `llvm/test/CodeGen/AMDGPU/*.ll`, leaving only those tests where its removal currently causes failures.
39 lines
1.4 KiB
LLVM
39 lines
1.4 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 %s -o - | FileCheck -check-prefixes=GCN %s
|
|
; xUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 %s -o - | FileCheck -check-prefix=GCN %s
|
|
|
|
; FIXME: GlobalISel does not work with bf16
|
|
|
|
declare bfloat @llvm.cos.bf16(bfloat) #0
|
|
|
|
define amdgpu_kernel void @cos_bf16_constant_4(ptr addrspace(1) %out) #1 {
|
|
; GCN-LABEL: cos_bf16_constant_4:
|
|
; GCN: ; %bb.0:
|
|
; GCN-NEXT: s_load_b64 s[0:1], s[4:5], 0x0
|
|
; GCN-NEXT: v_cos_bf16_e32 v0, 0x3f23
|
|
; GCN-NEXT: v_mov_b32_e32 v1, 0
|
|
; GCN-NEXT: s_wait_kmcnt 0x0
|
|
; GCN-NEXT: global_store_b16 v1, v0, s[0:1]
|
|
; GCN-NEXT: s_endpgm
|
|
%cos = call bfloat @llvm.cos.bf16(bfloat 4.0) #0
|
|
store bfloat %cos, ptr addrspace(1) %out, align 2
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_kernel void @cos_bf16_constant_100(ptr addrspace(1) %out) #1 {
|
|
; GCN-LABEL: cos_bf16_constant_100:
|
|
; GCN: ; %bb.0:
|
|
; GCN-NEXT: s_load_b64 s[0:1], s[4:5], 0x0
|
|
; GCN-NEXT: v_cos_bf16_e32 v0, 0x417f
|
|
; GCN-NEXT: v_mov_b32_e32 v1, 0
|
|
; GCN-NEXT: s_wait_kmcnt 0x0
|
|
; GCN-NEXT: global_store_b16 v1, v0, s[0:1]
|
|
; GCN-NEXT: s_endpgm
|
|
%cos = call bfloat @llvm.cos.bf16(bfloat 100.0) #0
|
|
store bfloat %cos, ptr addrspace(1) %out, align 2
|
|
ret void
|
|
}
|
|
|
|
attributes #0 = { nounwind readnone }
|
|
attributes #1 = { nounwind }
|