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.
28 lines
1.7 KiB
LLVM
28 lines
1.7 KiB
LLVM
; RUN: llc -global-isel=0 < %s -mtriple=amdgcn -mcpu=gfx600 | FileCheck %s -check-prefixes=GCN,SI
|
|
; RUN: llc -global-isel=0 < %s -mtriple=amdgcn -mcpu=gfx700 | FileCheck %s -check-prefixes=GCN,GCNX3
|
|
|
|
; GCN-LABEL: {{^}}tbuffer_raw_load_immoffs_x3:
|
|
; SI: tbuffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 format:[BUF_DATA_FORMAT_32_32_32_32,BUF_NUM_FORMAT_UINT] offset:42
|
|
; GCNX3: tbuffer_load_format_xyz {{v\[[0-9]+:[0-9]+\]}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 format:[BUF_DATA_FORMAT_32_32_32_32,BUF_NUM_FORMAT_UINT] offset:42
|
|
define amdgpu_vs <3 x float> @tbuffer_raw_load_immoffs_x3(<4 x i32> inreg) {
|
|
main_body:
|
|
%vdata = call <3 x i32> @llvm.amdgcn.raw.tbuffer.load.v3i32(<4 x i32> %0, i32 42, i32 0, i32 78, i32 0)
|
|
%vdata.f = bitcast <3 x i32> %vdata to <3 x float>
|
|
ret <3 x float> %vdata.f
|
|
}
|
|
|
|
|
|
; GCN-LABEL: {{^}}tbuffer_struct_load_immoffs_x3:
|
|
; GCN: v_mov_b32_e32 [[ZEROREG:v[0-9]+]], 0
|
|
; SI: tbuffer_load_format_xyzw {{v\[[0-9]+:[0-9]+\]}}, [[ZEROREG]], {{s\[[0-9]+:[0-9]+\]}}, 0 format:[BUF_DATA_FORMAT_32_32_32_32,BUF_NUM_FORMAT_UINT] idxen offset:42
|
|
; GCNX3: tbuffer_load_format_xyz {{v\[[0-9]+:[0-9]+\]}}, [[ZEROREG]], {{s\[[0-9]+:[0-9]+\]}}, 0 format:[BUF_DATA_FORMAT_32_32_32_32,BUF_NUM_FORMAT_UINT] idxen offset:42
|
|
define amdgpu_vs <3 x float> @tbuffer_struct_load_immoffs_x3(<4 x i32> inreg) {
|
|
main_body:
|
|
%vdata = call <3 x i32> @llvm.amdgcn.struct.tbuffer.load.v3i32(<4 x i32> %0, i32 0, i32 42, i32 0, i32 78, i32 0)
|
|
%vdata.f = bitcast <3 x i32> %vdata to <3 x float>
|
|
ret <3 x float> %vdata.f
|
|
}
|
|
|
|
declare <3 x i32> @llvm.amdgcn.raw.tbuffer.load.v3i32(<4 x i32>, i32, i32, i32, i32)
|
|
declare <3 x i32> @llvm.amdgcn.struct.tbuffer.load.v3i32(<4 x i32>, i32, i32, i32, i32, i32)
|