shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions 6 Packages Projects Releases Wiki Activity

[AMDGPU][test] Remove unsafe-fp-math uses (NFC) (#164609)

Browse Source 
Post cleanup for #164534.
This commit is contained in:
paperchalice 2025-10-23 09:45:54 +08:00 committed by GitHub
parent 1d9f9ad531
commit c2b2a347bf
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
14 changed files with 27 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
llvm/test/CodeGen/AMDGPU/combine_andor_with_cmps.ll
View File

@ -3385,7 +3385,7 @@ declare half @llvm.canonicalize.f16(half)
declare <2 x half> @llvm.canonicalize.v2f16(<2 x half>)
attributes #0 = { nounwind "amdgpu-ieee"="false" }
attributes #1 = { nounwind "unsafe-fp-math"="true" "no-nans-fp-math"="true" }
attributes #1 = { nounwind "no-nans-fp-math"="true" }
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; GFX11NONANS-FAKE16: {{.*}}
; GFX11NONANS-TRUE16: {{.*}}

7
llvm/test/CodeGen/AMDGPU/fdiv.f64.ll
View File

@ -127,7 +127,7 @@ define amdgpu_kernel void @s_fdiv_v4f64(ptr addrspace(1) %out, <4 x double> %num
; GCN-LABEL: {{^}}div_fast_2_x_pat_f64:
; GCN: v_mul_f64 [[MUL:v\[[0-9]+:[0-9]+\]]], s{{\[[0-9]+:[0-9]+\]}}, 0.5
; GCN: buffer_store_dwordx2 [[MUL]]
define amdgpu_kernel void @div_fast_2_x_pat_f64(ptr addrspace(1) %out) #1 {
define amdgpu_kernel void @div_fast_2_x_pat_f64(ptr addrspace(1) %out) #0 {
%x = load double, ptr addrspace(1) poison
%rcp = fdiv fast double %x, 2.0
store double %rcp, ptr addrspace(1) %out, align 4
@ -139,7 +139,7 @@ define amdgpu_kernel void @div_fast_2_x_pat_f64(ptr addrspace(1) %out) #1 {
; GCN-DAG: v_mov_b32_e32 v[[K_HI:[0-9]+]], 0x3fb99999
; GCN: v_mul_f64 [[MUL:v\[[0-9]+:[0-9]+\]]], s{{\[[0-9]+:[0-9]+\]}}, v[[[K_LO]]:[[K_HI]]]
; GCN: buffer_store_dwordx2 [[MUL]]
define amdgpu_kernel void @div_fast_k_x_pat_f64(ptr addrspace(1) %out) #1 {
define amdgpu_kernel void @div_fast_k_x_pat_f64(ptr addrspace(1) %out) #0 {
%x = load double, ptr addrspace(1) poison
%rcp = fdiv fast double %x, 10.0
store double %rcp, ptr addrspace(1) %out, align 4
@ -151,7 +151,7 @@ define amdgpu_kernel void @div_fast_k_x_pat_f64(ptr addrspace(1) %out) #1 {
; GCN-DAG: v_mov_b32_e32 v[[K_HI:[0-9]+]], 0xbfb99999
; GCN: v_mul_f64 [[MUL:v\[[0-9]+:[0-9]+\]]], s{{\[[0-9]+:[0-9]+\]}}, v[[[K_LO]]:[[K_HI]]]
; GCN: buffer_store_dwordx2 [[MUL]]
define amdgpu_kernel void @div_fast_neg_k_x_pat_f64(ptr addrspace(1) %out) #1 {
define amdgpu_kernel void @div_fast_neg_k_x_pat_f64(ptr addrspace(1) %out) #0 {
%x = load double, ptr addrspace(1) poison
%rcp = fdiv fast double %x, -10.0
store double %rcp, ptr addrspace(1) %out, align 4
@ -159,4 +159,3 @@ define amdgpu_kernel void @div_fast_neg_k_x_pat_f64(ptr addrspace(1) %out) #1 {
}
attributes #0 = { nounwind }
attributes #1 = { nounwind "unsafe-fp-math"="true" }

2
llvm/test/CodeGen/AMDGPU/fmad-formation-fmul-distribute-denormal-mode.ll
View File

@ -284,4 +284,4 @@ define <2 x float> @unsafe_fast_fmul_fsub_ditribute_post_legalize(float %arg0, <
ret <2 x float> %tmp1
}
attributes #0 = { "no-infs-fp-math"="true" "unsafe-fp-math"="true" }
attributes #0 = { "no-infs-fp-math"="true" }

4
llvm/test/CodeGen/AMDGPU/fmed3.bf16.ll
View File

@ -219,8 +219,8 @@ define <2 x bfloat> @v_test_fmed3_r_i_i_v2bf16_minimumnum_maximumnum(<2 x bfloat
}
attributes #0 = { nounwind readnone }
attributes #1 = { nounwind "unsafe-fp-math"="false" "no-nans-fp-math"="false" }
attributes #2 = { nounwind "unsafe-fp-math"="false" "no-nans-fp-math"="true" }
attributes #1 = { nounwind "no-nans-fp-math"="false" }
attributes #2 = { nounwind "no-nans-fp-math"="true" }
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; GFX11: {{.*}}
; GFX11-SDAG: {{.*}}

2
llvm/test/CodeGen/AMDGPU/fmed3.ll
View File

@ -8905,4 +8905,4 @@ declare half @llvm.minnum.f16(half, half) #0
declare half @llvm.maxnum.f16(half, half) #0
attributes #0 = { nounwind readnone }
attributes #2 = { nounwind "unsafe-fp-math"="false" "no-nans-fp-math"="true" }
attributes #2 = { nounwind "no-nans-fp-math"="true" }

2
llvm/test/CodeGen/AMDGPU/fneg-combines.legal.f16.ll
View File

@ -159,7 +159,7 @@ declare half @llvm.amdgcn.interp.p2.f16(float, float, i32, i32, i1, i32) #0
attributes #0 = { nounwind "denormal-fp-math-f32"="preserve-sign,preserve-sign" }
attributes #1 = { nounwind readnone }
attributes #2 = { nounwind "unsafe-fp-math"="true" }
attributes #2 = { nounwind }
attributes #3 = { nounwind "no-signed-zeros-fp-math"="true" }
attributes #4 = { nounwind "amdgpu-ieee"="false" "denormal-fp-math-f32"="preserve-sign,preserve-sign" }
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:

2
llvm/test/CodeGen/AMDGPU/fneg-combines.ll
View File

@ -8006,7 +8006,7 @@ declare float @llvm.amdgcn.interp.p2(float, float, i32, i32, i32) #0
attributes #0 = { nounwind "denormal-fp-math-f32"="preserve-sign,preserve-sign" }
attributes #1 = { nounwind readnone }
attributes #2 = { nounwind "unsafe-fp-math"="true" }
attributes #2 = { nounwind }
attributes #3 = { nounwind "no-signed-zeros-fp-math"="true" }
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; GCN-NSZ: {{.*}}

11
llvm/test/CodeGen/AMDGPU/frem.ll
View File

@ -2048,7 +2048,7 @@ define amdgpu_kernel void @unsafe_frem_f16(ptr addrspace(1) %out, ptr addrspace(
; GFX1200-FAKE16-NEXT: v_fmac_f16_e32 v1, v3, v2
; GFX1200-FAKE16-NEXT: global_store_b16 v0, v1, s[0:1]
; GFX1200-FAKE16-NEXT: s_endpgm
ptr addrspace(1) %in2) #1 {
ptr addrspace(1) %in2) #0 {
%gep2 = getelementptr half, ptr addrspace(1) %in2, i32 4
%r0 = load half, ptr addrspace(1) %in1, align 4
%r1 = load half, ptr addrspace(1) %gep2, align 4
@ -3417,7 +3417,7 @@ define amdgpu_kernel void @unsafe_frem_f32(ptr addrspace(1) %out, ptr addrspace(
; GFX1200-NEXT: v_fmac_f32_e32 v1, v3, v2
; GFX1200-NEXT: global_store_b32 v0, v1, s[0:1]
; GFX1200-NEXT: s_endpgm
ptr addrspace(1) %in2) #1 {
ptr addrspace(1) %in2) #0 {
%gep2 = getelementptr float, ptr addrspace(1) %in2, i32 4
%r0 = load float, ptr addrspace(1) %in1, align 4
%r1 = load float, ptr addrspace(1) %gep2, align 4
@ -4821,7 +4821,7 @@ define amdgpu_kernel void @unsafe_frem_f64(ptr addrspace(1) %out, ptr addrspace(
; GFX1200-NEXT: v_fma_f64 v[0:1], -v[4:5], v[2:3], v[0:1]
; GFX1200-NEXT: global_store_b64 v12, v[0:1], s[0:1]
; GFX1200-NEXT: s_endpgm
ptr addrspace(1) %in2) #1 {
ptr addrspace(1) %in2) #0 {
%r0 = load double, ptr addrspace(1) %in1, align 8
%r1 = load double, ptr addrspace(1) %in2, align 8
%r2 = frem afn double %r0, %r1
@ -18918,7 +18918,4 @@ define amdgpu_kernel void @frem_v2f64_const(ptr addrspace(1) %out) #0 {
attributes #0 = { nounwind "unsafe-fp-math"="false" "denormal-fp-math-f32"="preserve-sign,preserve-sign" }
attributes #1 = { nounwind "unsafe-fp-math"="true" "denormal-fp-math-f32"="preserve-sign,preserve-sign" }
attributes #0 = { nounwind "denormal-fp-math-f32"="preserve-sign,preserve-sign" }

3
llvm/test/CodeGen/AMDGPU/fsqrt.f64.ll
View File

@ -2870,7 +2870,7 @@ define double @v_sqrt_f64__enough_unsafe_attrs(double %x) #3 {
ret double %result
}
define double @v_sqrt_f64__unsafe_attr(double %x) #4 {
define double @v_sqrt_f64__unsafe_attr(double %x) {
; GFX6-SDAG-LABEL: v_sqrt_f64__unsafe_attr:
; GFX6-SDAG: ; %bb.0:
; GFX6-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
@ -3449,7 +3449,6 @@ declare i32 @llvm.amdgcn.readfirstlane(i32) #1
attributes #0 = { nocallback nofree nosync nounwind speculatable willreturn memory(none) }
attributes #1 = { convergent nounwind willreturn memory(none) }
attributes #3 = { "no-nans-fp-math"="true" "no-infs-fp-math"="true" }
attributes #4 = { "unsafe-fp-math"="true" }
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; GFX6: {{.*}}
; GFX8: {{.*}}

2
llvm/test/CodeGen/AMDGPU/fsqrt.r600.ll
View File

@ -239,4 +239,4 @@ declare <2 x float> @llvm.sqrt.v2f32(<2 x float> %in) #0
declare <4 x float> @llvm.sqrt.v4f32(<4 x float> %in) #0
attributes #0 = { nounwind readnone }
attributes #1 = { nounwind "unsafe-fp-math"="true" }
attributes #1 = { nounwind }

16
llvm/test/CodeGen/AMDGPU/inline-attr.ll
View File

@ -36,18 +36,18 @@ entry:
ret void
}
attributes #0 = { nounwind "uniform-work-group-size"="false" "unsafe-fp-math"="true"}
attributes #1 = { nounwind "less-precise-fpmad"="true" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "unsafe-fp-math"="true" }
attributes #0 = { nounwind "uniform-work-group-size"="false"}
attributes #1 = { nounwind "less-precise-fpmad"="true" "no-infs-fp-math"="true" "no-nans-fp-math"="true" }
;.
; UNSAFE: attributes #[[ATTR0]] = { nounwind "uniform-work-group-size"="false" "unsafe-fp-math"="true" }
; UNSAFE: attributes #[[ATTR1]] = { nounwind "less-precise-fpmad"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "uniform-work-group-size"="false" "unsafe-fp-math"="true" }
; UNSAFE: attributes #[[ATTR0]] = { nounwind "uniform-work-group-size"="false" }
; UNSAFE: attributes #[[ATTR1]] = { nounwind "less-precise-fpmad"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "uniform-work-group-size"="false" }
;.
; NONANS: attributes #[[ATTR0]] = { nounwind "no-nans-fp-math"="true" "uniform-work-group-size"="false" "unsafe-fp-math"="true" }
; NONANS: attributes #[[ATTR1]] = { nounwind "less-precise-fpmad"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="true" "uniform-work-group-size"="false" "unsafe-fp-math"="true" }
; NONANS: attributes #[[ATTR0]] = { nounwind "no-nans-fp-math"="true" "uniform-work-group-size"="false" }
; NONANS: attributes #[[ATTR1]] = { nounwind "less-precise-fpmad"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="true" "uniform-work-group-size"="false" }
;.
; NOINFS: attributes #[[ATTR0]] = { nounwind "no-infs-fp-math"="true" "uniform-work-group-size"="false" "unsafe-fp-math"="true" }
; NOINFS: attributes #[[ATTR1]] = { nounwind "less-precise-fpmad"="false" "no-infs-fp-math"="true" "no-nans-fp-math"="false" "uniform-work-group-size"="false" "unsafe-fp-math"="true" }
; NOINFS: attributes #[[ATTR0]] = { nounwind "no-infs-fp-math"="true" "uniform-work-group-size"="false" }
; NOINFS: attributes #[[ATTR1]] = { nounwind "less-precise-fpmad"="false" "no-infs-fp-math"="true" "no-nans-fp-math"="false" "uniform-work-group-size"="false" }
;.
; UNSAFE: [[META0]] = !{}
;.

2
llvm/test/CodeGen/AMDGPU/llvm.exp2.ll
View File

@ -1485,7 +1485,7 @@ define float @v_exp2_f32_fast(float %in) {
ret float %result
}
define float @v_exp2_f32_unsafe_math_attr(float %in) "unsafe-fp-math"="true" {
define float @v_exp2_f32_unsafe_math_attr(float %in) {
; SI-SDAG-LABEL: v_exp2_f32_unsafe_math_attr:
; SI-SDAG: ; %bb.0:
; SI-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)

2
llvm/test/CodeGen/AMDGPU/llvm.log2.ll
View File

@ -1907,7 +1907,7 @@ define float @v_log2_f32_fast(float %in) {
ret float %result
}
define float @v_log2_f32_unsafe_math_attr(float %in) "unsafe-fp-math"="true" {
define float @v_log2_f32_unsafe_math_attr(float %in) {
; SI-SDAG-LABEL: v_log2_f32_unsafe_math_attr:
; SI-SDAG: ; %bb.0:
; SI-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)

2
llvm/test/CodeGen/AMDGPU/minmax.ll
View File

@ -1296,4 +1296,4 @@ declare half @llvm.minnum.f16(half, half)
declare half @llvm.maxnum.f16(half, half)
declare float @llvm.minnum.f32(float, float)
declare float @llvm.maxnum.f32(float, float)
attributes #0 = { nounwind "unsafe-fp-math"="false" "no-nans-fp-math"="true" }
attributes #0 = { nounwind "no-nans-fp-math"="true" }