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.
52 lines
2.4 KiB
LLVM
52 lines
2.4 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -global-isel=0 -mtriple=amdgcn -mcpu=gfx1200 | FileCheck %s -check-prefixes=GFX12,GFX12-SDAG
|
|
; RUN: llc < %s -global-isel=1 -mtriple=amdgcn -mcpu=gfx1200 | FileCheck %s -check-prefixes=GFX12,GFX12-GISEL
|
|
|
|
declare float @llvm.amdgcn.global.atomic.fmin.num.f32.p1.f32(ptr addrspace(1) %ptr, float %data)
|
|
declare float @llvm.amdgcn.global.atomic.fmax.num.f32.p1.f32(ptr addrspace(1) %ptr, float %data)
|
|
|
|
define amdgpu_cs void @global_atomic_fmin_num_f32_noret(ptr addrspace(1) %ptr, float %data) {
|
|
; GFX12-LABEL: global_atomic_fmin_num_f32_noret:
|
|
; GFX12: ; %bb.0:
|
|
; GFX12-NEXT: global_atomic_min_num_f32 v[0:1], v2, off
|
|
; GFX12-NEXT: s_endpgm
|
|
%ret = call float @llvm.amdgcn.global.atomic.fmin.num.f32.p1.f32(ptr addrspace(1) %ptr, float %data)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_cs void @global_atomic_fmax_num_f32_noret(ptr addrspace(1) %ptr, float %data) {
|
|
; GFX12-LABEL: global_atomic_fmax_num_f32_noret:
|
|
; GFX12: ; %bb.0:
|
|
; GFX12-NEXT: global_atomic_max_num_f32 v[0:1], v2, off
|
|
; GFX12-NEXT: s_endpgm
|
|
%ret = call float @llvm.amdgcn.global.atomic.fmax.num.f32.p1.f32(ptr addrspace(1) %ptr, float %data)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_cs void @global_atomic_fmax_num_f32_rtn(ptr addrspace(1) %ptr, float %data, ptr addrspace(1) %out) {
|
|
; GFX12-LABEL: global_atomic_fmax_num_f32_rtn:
|
|
; GFX12: ; %bb.0:
|
|
; GFX12-NEXT: global_atomic_max_num_f32 v0, v[0:1], v2, off th:TH_ATOMIC_RETURN
|
|
; GFX12-NEXT: s_wait_loadcnt 0x0
|
|
; GFX12-NEXT: global_store_b32 v[3:4], v0, off
|
|
; GFX12-NEXT: s_endpgm
|
|
%ret = call float @llvm.amdgcn.global.atomic.fmax.num.f32.p1.f32(ptr addrspace(1) %ptr, float %data)
|
|
store float %ret, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_cs void @global_atomic_fmin_num_f32_rtn(ptr addrspace(1) %ptr, float %data, ptr addrspace(1) %out) {
|
|
; GFX12-LABEL: global_atomic_fmin_num_f32_rtn:
|
|
; GFX12: ; %bb.0:
|
|
; GFX12-NEXT: global_atomic_min_num_f32 v0, v[0:1], v2, off th:TH_ATOMIC_RETURN
|
|
; GFX12-NEXT: s_wait_loadcnt 0x0
|
|
; GFX12-NEXT: global_store_b32 v[3:4], v0, off
|
|
; GFX12-NEXT: s_endpgm
|
|
%ret = call float @llvm.amdgcn.global.atomic.fmin.num.f32.p1.f32(ptr addrspace(1) %ptr, float %data)
|
|
store float %ret, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
|
|
; GFX12-GISEL: {{.*}}
|
|
; GFX12-SDAG: {{.*}}
|