[llvm] [NFC] #155740 post cleanup (PR #155966)
via llvm-commits
llvm-commits at lists.llvm.org
Thu Aug 28 21:14:09 PDT 2025
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-directx
Author: None (paperchalice)
<details>
<summary>Changes</summary>
Remove all "approx-func-fp-math" in tests.
---
Patch is 49.69 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/155966.diff
17 Files Affected:
- (modified) llvm/test/CodeGen/AMDGPU/fsqrt.f64.ll (+1-118)
- (modified) llvm/test/CodeGen/AMDGPU/llvm.exp2.ll (-98)
- (modified) llvm/test/CodeGen/AMDGPU/llvm.log2.ll (-123)
- (modified) llvm/test/CodeGen/DirectX/ForwardHandleAccesses/cbuffer-access.ll (+1-1)
- (modified) llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-cs.ll (+2-2)
- (modified) llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-lib.ll (+2-2)
- (modified) llvm/test/CodeGen/DirectX/ShaderFlags/lib-entry-attr-error.ll (+2-2)
- (modified) llvm/test/CodeGen/Hexagon/loopIdiom.ll (+1-1)
- (modified) llvm/test/CodeGen/PowerPC/lower-scalar-mass-afn.ll (+79-81)
- (modified) llvm/test/CodeGen/PowerPC/pow-025-075-intrinsic-scalar-mass-fast.ll (+1-1)
- (modified) llvm/test/CodeGen/PowerPC/pow-025-075-nointrinsic-scalar-mass-fast.ll (+1-1)
- (modified) llvm/test/CodeGen/SPIRV/hlsl-resources/StructuredBuffer.ll (+1-1)
- (modified) llvm/test/CodeGen/SPIRV/hlsl-resources/spirv.layout.type.ll (+1-1)
- (modified) llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll (+1-1)
- (modified) llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll (+2-2)
- (modified) llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll (+1-1)
- (modified) llvm/test/Transforms/PhaseOrdering/ARM/arm_mean_q7.ll (+1-1)
``````````diff
diff --git a/llvm/test/CodeGen/AMDGPU/fsqrt.f64.ll b/llvm/test/CodeGen/AMDGPU/fsqrt.f64.ll
index cffa287dd91f5..76e15eed08cc2 100644
--- a/llvm/test/CodeGen/AMDGPU/fsqrt.f64.ll
+++ b/llvm/test/CodeGen/AMDGPU/fsqrt.f64.ll
@@ -2762,122 +2762,6 @@ define double @v_sqrt_f64_afn_nnan_ninf_nsz(double %x) {
ret double %result
}
-define double @v_sqrt_f64__approx_func_fp_math(double %x) #2 {
-; GFX6-SDAG-LABEL: v_sqrt_f64__approx_func_fp_math:
-; GFX6-SDAG: ; %bb.0:
-; GFX6-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX6-SDAG-NEXT: s_mov_b32 s4, 0
-; GFX6-SDAG-NEXT: s_brev_b32 s5, 8
-; GFX6-SDAG-NEXT: v_cmp_gt_f64_e32 vcc, s[4:5], v[0:1]
-; GFX6-SDAG-NEXT: v_mov_b32_e32 v2, 0x100
-; GFX6-SDAG-NEXT: v_cndmask_b32_e32 v2, 0, v2, vcc
-; GFX6-SDAG-NEXT: v_ldexp_f64 v[0:1], v[0:1], v2
-; GFX6-SDAG-NEXT: v_rsq_f64_e32 v[2:3], v[0:1]
-; GFX6-SDAG-NEXT: v_mul_f64 v[4:5], v[0:1], v[2:3]
-; GFX6-SDAG-NEXT: v_mul_f64 v[2:3], v[2:3], 0.5
-; GFX6-SDAG-NEXT: v_fma_f64 v[6:7], -v[2:3], v[4:5], 0.5
-; GFX6-SDAG-NEXT: v_fma_f64 v[4:5], v[4:5], v[6:7], v[4:5]
-; GFX6-SDAG-NEXT: v_fma_f64 v[2:3], v[2:3], v[6:7], v[2:3]
-; GFX6-SDAG-NEXT: v_fma_f64 v[6:7], -v[4:5], v[4:5], v[0:1]
-; GFX6-SDAG-NEXT: v_fma_f64 v[4:5], v[6:7], v[2:3], v[4:5]
-; GFX6-SDAG-NEXT: v_fma_f64 v[6:7], -v[4:5], v[4:5], v[0:1]
-; GFX6-SDAG-NEXT: v_fma_f64 v[2:3], v[6:7], v[2:3], v[4:5]
-; GFX6-SDAG-NEXT: v_mov_b32_e32 v4, 0xffffff80
-; GFX6-SDAG-NEXT: v_mov_b32_e32 v5, 0x260
-; GFX6-SDAG-NEXT: v_cndmask_b32_e32 v4, 0, v4, vcc
-; GFX6-SDAG-NEXT: v_cmp_class_f64_e32 vcc, v[0:1], v5
-; GFX6-SDAG-NEXT: v_ldexp_f64 v[2:3], v[2:3], v4
-; GFX6-SDAG-NEXT: v_cndmask_b32_e32 v0, v2, v0, vcc
-; GFX6-SDAG-NEXT: v_cndmask_b32_e32 v1, v3, v1, vcc
-; GFX6-SDAG-NEXT: s_setpc_b64 s[30:31]
-;
-; GFX8-SDAG-LABEL: v_sqrt_f64__approx_func_fp_math:
-; GFX8-SDAG: ; %bb.0:
-; GFX8-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX8-SDAG-NEXT: s_mov_b32 s4, 0
-; GFX8-SDAG-NEXT: s_brev_b32 s5, 8
-; GFX8-SDAG-NEXT: v_cmp_gt_f64_e32 vcc, s[4:5], v[0:1]
-; GFX8-SDAG-NEXT: v_mov_b32_e32 v2, 0x100
-; GFX8-SDAG-NEXT: v_cndmask_b32_e32 v2, 0, v2, vcc
-; GFX8-SDAG-NEXT: v_ldexp_f64 v[0:1], v[0:1], v2
-; GFX8-SDAG-NEXT: v_rsq_f64_e32 v[2:3], v[0:1]
-; GFX8-SDAG-NEXT: v_mul_f64 v[4:5], v[0:1], v[2:3]
-; GFX8-SDAG-NEXT: v_mul_f64 v[2:3], v[2:3], 0.5
-; GFX8-SDAG-NEXT: v_fma_f64 v[6:7], -v[2:3], v[4:5], 0.5
-; GFX8-SDAG-NEXT: v_fma_f64 v[4:5], v[4:5], v[6:7], v[4:5]
-; GFX8-SDAG-NEXT: v_fma_f64 v[2:3], v[2:3], v[6:7], v[2:3]
-; GFX8-SDAG-NEXT: v_fma_f64 v[6:7], -v[4:5], v[4:5], v[0:1]
-; GFX8-SDAG-NEXT: v_fma_f64 v[4:5], v[6:7], v[2:3], v[4:5]
-; GFX8-SDAG-NEXT: v_fma_f64 v[6:7], -v[4:5], v[4:5], v[0:1]
-; GFX8-SDAG-NEXT: v_fma_f64 v[2:3], v[6:7], v[2:3], v[4:5]
-; GFX8-SDAG-NEXT: v_mov_b32_e32 v4, 0xffffff80
-; GFX8-SDAG-NEXT: v_mov_b32_e32 v5, 0x260
-; GFX8-SDAG-NEXT: v_cndmask_b32_e32 v4, 0, v4, vcc
-; GFX8-SDAG-NEXT: v_cmp_class_f64_e32 vcc, v[0:1], v5
-; GFX8-SDAG-NEXT: v_ldexp_f64 v[2:3], v[2:3], v4
-; GFX8-SDAG-NEXT: v_cndmask_b32_e32 v0, v2, v0, vcc
-; GFX8-SDAG-NEXT: v_cndmask_b32_e32 v1, v3, v1, vcc
-; GFX8-SDAG-NEXT: s_setpc_b64 s[30:31]
-;
-; GFX6-GISEL-LABEL: v_sqrt_f64__approx_func_fp_math:
-; GFX6-GISEL: ; %bb.0:
-; GFX6-GISEL-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX6-GISEL-NEXT: v_mov_b32_e32 v2, 0
-; GFX6-GISEL-NEXT: v_bfrev_b32_e32 v3, 8
-; GFX6-GISEL-NEXT: v_cmp_lt_f64_e32 vcc, v[0:1], v[2:3]
-; GFX6-GISEL-NEXT: v_cndmask_b32_e64 v2, 0, 1, vcc
-; GFX6-GISEL-NEXT: v_lshlrev_b32_e32 v2, 8, v2
-; GFX6-GISEL-NEXT: v_ldexp_f64 v[0:1], v[0:1], v2
-; GFX6-GISEL-NEXT: v_rsq_f64_e32 v[2:3], v[0:1]
-; GFX6-GISEL-NEXT: v_mul_f64 v[4:5], v[2:3], 0.5
-; GFX6-GISEL-NEXT: v_mul_f64 v[2:3], v[0:1], v[2:3]
-; GFX6-GISEL-NEXT: v_fma_f64 v[6:7], -v[4:5], v[2:3], 0.5
-; GFX6-GISEL-NEXT: v_fma_f64 v[2:3], v[2:3], v[6:7], v[2:3]
-; GFX6-GISEL-NEXT: v_fma_f64 v[4:5], v[4:5], v[6:7], v[4:5]
-; GFX6-GISEL-NEXT: v_fma_f64 v[6:7], -v[2:3], v[2:3], v[0:1]
-; GFX6-GISEL-NEXT: v_fma_f64 v[2:3], v[6:7], v[4:5], v[2:3]
-; GFX6-GISEL-NEXT: v_fma_f64 v[6:7], -v[2:3], v[2:3], v[0:1]
-; GFX6-GISEL-NEXT: v_fma_f64 v[2:3], v[6:7], v[4:5], v[2:3]
-; GFX6-GISEL-NEXT: v_mov_b32_e32 v4, 0xffffff80
-; GFX6-GISEL-NEXT: v_mov_b32_e32 v5, 0x260
-; GFX6-GISEL-NEXT: v_cndmask_b32_e32 v4, 0, v4, vcc
-; GFX6-GISEL-NEXT: v_cmp_class_f64_e32 vcc, v[0:1], v5
-; GFX6-GISEL-NEXT: v_ldexp_f64 v[2:3], v[2:3], v4
-; GFX6-GISEL-NEXT: v_cndmask_b32_e32 v0, v2, v0, vcc
-; GFX6-GISEL-NEXT: v_cndmask_b32_e32 v1, v3, v1, vcc
-; GFX6-GISEL-NEXT: s_setpc_b64 s[30:31]
-;
-; GFX8-GISEL-LABEL: v_sqrt_f64__approx_func_fp_math:
-; GFX8-GISEL: ; %bb.0:
-; GFX8-GISEL-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX8-GISEL-NEXT: v_mov_b32_e32 v2, 0
-; GFX8-GISEL-NEXT: v_bfrev_b32_e32 v3, 8
-; GFX8-GISEL-NEXT: v_cmp_lt_f64_e32 vcc, v[0:1], v[2:3]
-; GFX8-GISEL-NEXT: v_cndmask_b32_e64 v2, 0, 1, vcc
-; GFX8-GISEL-NEXT: v_lshlrev_b32_e32 v2, 8, v2
-; GFX8-GISEL-NEXT: v_ldexp_f64 v[0:1], v[0:1], v2
-; GFX8-GISEL-NEXT: v_rsq_f64_e32 v[2:3], v[0:1]
-; GFX8-GISEL-NEXT: v_mul_f64 v[4:5], v[2:3], 0.5
-; GFX8-GISEL-NEXT: v_mul_f64 v[2:3], v[0:1], v[2:3]
-; GFX8-GISEL-NEXT: v_fma_f64 v[6:7], -v[4:5], v[2:3], 0.5
-; GFX8-GISEL-NEXT: v_fma_f64 v[2:3], v[2:3], v[6:7], v[2:3]
-; GFX8-GISEL-NEXT: v_fma_f64 v[4:5], v[4:5], v[6:7], v[4:5]
-; GFX8-GISEL-NEXT: v_fma_f64 v[6:7], -v[2:3], v[2:3], v[0:1]
-; GFX8-GISEL-NEXT: v_fma_f64 v[2:3], v[6:7], v[4:5], v[2:3]
-; GFX8-GISEL-NEXT: v_fma_f64 v[6:7], -v[2:3], v[2:3], v[0:1]
-; GFX8-GISEL-NEXT: v_fma_f64 v[2:3], v[6:7], v[4:5], v[2:3]
-; GFX8-GISEL-NEXT: v_mov_b32_e32 v4, 0xffffff80
-; GFX8-GISEL-NEXT: v_mov_b32_e32 v5, 0x260
-; GFX8-GISEL-NEXT: v_cndmask_b32_e32 v4, 0, v4, vcc
-; GFX8-GISEL-NEXT: v_cmp_class_f64_e32 vcc, v[0:1], v5
-; GFX8-GISEL-NEXT: v_ldexp_f64 v[2:3], v[2:3], v4
-; GFX8-GISEL-NEXT: v_cndmask_b32_e32 v0, v2, v0, vcc
-; GFX8-GISEL-NEXT: v_cndmask_b32_e32 v1, v3, v1, vcc
-; GFX8-GISEL-NEXT: s_setpc_b64 s[30:31]
- %result = call nsz double @llvm.sqrt.f64(double %x)
- ret double %result
-}
-
define double @v_sqrt_f64__enough_unsafe_attrs(double %x) #3 {
; GFX6-SDAG-LABEL: v_sqrt_f64__enough_unsafe_attrs:
; GFX6-SDAG: ; %bb.0:
@@ -3580,8 +3464,7 @@ 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 #2 = { "approx-func-fp-math"="true" }
-attributes #3 = { "approx-func-fp-math"="true" "no-nans-fp-math"="true" "no-infs-fp-math"="true" }
+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: {{.*}}
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.exp2.ll b/llvm/test/CodeGen/AMDGPU/llvm.exp2.ll
index e71ea505caea1..883db20a867b3 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.exp2.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.exp2.ll
@@ -1583,104 +1583,6 @@ define float @v_exp2_f32_unsafe_math_attr(float %in) "unsafe-fp-math"="true" {
ret float %result
}
-define float @v_exp2_f32_approx_fn_attr(float %in) "approx-func-fp-math"="true" {
-; SI-SDAG-LABEL: v_exp2_f32_approx_fn_attr:
-; SI-SDAG: ; %bb.0:
-; SI-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; SI-SDAG-NEXT: s_mov_b32 s4, 0xc2fc0000
-; SI-SDAG-NEXT: v_cmp_gt_f32_e32 vcc, s4, v0
-; SI-SDAG-NEXT: v_mov_b32_e32 v2, 0x42800000
-; SI-SDAG-NEXT: v_cndmask_b32_e32 v2, 0, v2, vcc
-; SI-SDAG-NEXT: v_add_f32_e32 v0, v0, v2
-; SI-SDAG-NEXT: v_exp_f32_e32 v0, v0
-; SI-SDAG-NEXT: v_not_b32_e32 v1, 63
-; SI-SDAG-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; SI-SDAG-NEXT: v_ldexp_f32_e32 v0, v0, v1
-; SI-SDAG-NEXT: s_setpc_b64 s[30:31]
-;
-; SI-GISEL-LABEL: v_exp2_f32_approx_fn_attr:
-; SI-GISEL: ; %bb.0:
-; SI-GISEL-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; SI-GISEL-NEXT: v_mov_b32_e32 v1, 0xc2fc0000
-; SI-GISEL-NEXT: v_mov_b32_e32 v2, 0x42800000
-; SI-GISEL-NEXT: v_cmp_lt_f32_e32 vcc, v0, v1
-; SI-GISEL-NEXT: v_cndmask_b32_e32 v1, 0, v2, vcc
-; SI-GISEL-NEXT: v_add_f32_e32 v0, v0, v1
-; SI-GISEL-NEXT: v_exp_f32_e32 v0, v0
-; SI-GISEL-NEXT: v_not_b32_e32 v1, 63
-; SI-GISEL-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; SI-GISEL-NEXT: v_ldexp_f32_e32 v0, v0, v1
-; SI-GISEL-NEXT: s_setpc_b64 s[30:31]
-;
-; VI-SDAG-LABEL: v_exp2_f32_approx_fn_attr:
-; VI-SDAG: ; %bb.0:
-; VI-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; VI-SDAG-NEXT: s_mov_b32 s4, 0xc2fc0000
-; VI-SDAG-NEXT: v_cmp_gt_f32_e32 vcc, s4, v0
-; VI-SDAG-NEXT: v_mov_b32_e32 v2, 0x42800000
-; VI-SDAG-NEXT: v_cndmask_b32_e32 v2, 0, v2, vcc
-; VI-SDAG-NEXT: v_add_f32_e32 v0, v0, v2
-; VI-SDAG-NEXT: v_exp_f32_e32 v0, v0
-; VI-SDAG-NEXT: v_not_b32_e32 v1, 63
-; VI-SDAG-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; VI-SDAG-NEXT: v_ldexp_f32 v0, v0, v1
-; VI-SDAG-NEXT: s_setpc_b64 s[30:31]
-;
-; VI-GISEL-LABEL: v_exp2_f32_approx_fn_attr:
-; VI-GISEL: ; %bb.0:
-; VI-GISEL-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; VI-GISEL-NEXT: v_mov_b32_e32 v1, 0xc2fc0000
-; VI-GISEL-NEXT: v_mov_b32_e32 v2, 0x42800000
-; VI-GISEL-NEXT: v_cmp_lt_f32_e32 vcc, v0, v1
-; VI-GISEL-NEXT: v_cndmask_b32_e32 v1, 0, v2, vcc
-; VI-GISEL-NEXT: v_add_f32_e32 v0, v0, v1
-; VI-GISEL-NEXT: v_exp_f32_e32 v0, v0
-; VI-GISEL-NEXT: v_not_b32_e32 v1, 63
-; VI-GISEL-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; VI-GISEL-NEXT: v_ldexp_f32 v0, v0, v1
-; VI-GISEL-NEXT: s_setpc_b64 s[30:31]
-;
-; GFX900-SDAG-LABEL: v_exp2_f32_approx_fn_attr:
-; GFX900-SDAG: ; %bb.0:
-; GFX900-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX900-SDAG-NEXT: s_mov_b32 s4, 0xc2fc0000
-; GFX900-SDAG-NEXT: v_cmp_gt_f32_e32 vcc, s4, v0
-; GFX900-SDAG-NEXT: v_mov_b32_e32 v2, 0x42800000
-; GFX900-SDAG-NEXT: v_cndmask_b32_e32 v2, 0, v2, vcc
-; GFX900-SDAG-NEXT: v_add_f32_e32 v0, v0, v2
-; GFX900-SDAG-NEXT: v_exp_f32_e32 v0, v0
-; GFX900-SDAG-NEXT: v_not_b32_e32 v1, 63
-; GFX900-SDAG-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; GFX900-SDAG-NEXT: v_ldexp_f32 v0, v0, v1
-; GFX900-SDAG-NEXT: s_setpc_b64 s[30:31]
-;
-; GFX900-GISEL-LABEL: v_exp2_f32_approx_fn_attr:
-; GFX900-GISEL: ; %bb.0:
-; GFX900-GISEL-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX900-GISEL-NEXT: v_mov_b32_e32 v1, 0xc2fc0000
-; GFX900-GISEL-NEXT: v_mov_b32_e32 v2, 0x42800000
-; GFX900-GISEL-NEXT: v_cmp_lt_f32_e32 vcc, v0, v1
-; GFX900-GISEL-NEXT: v_cndmask_b32_e32 v1, 0, v2, vcc
-; GFX900-GISEL-NEXT: v_add_f32_e32 v0, v0, v1
-; GFX900-GISEL-NEXT: v_exp_f32_e32 v0, v0
-; GFX900-GISEL-NEXT: v_not_b32_e32 v1, 63
-; GFX900-GISEL-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; GFX900-GISEL-NEXT: v_ldexp_f32 v0, v0, v1
-; GFX900-GISEL-NEXT: s_setpc_b64 s[30:31]
-;
-; R600-LABEL: v_exp2_f32_approx_fn_attr:
-; R600: ; %bb.0:
-; R600-NEXT: CF_END
-; R600-NEXT: PAD
-;
-; CM-LABEL: v_exp2_f32_approx_fn_attr:
-; CM: ; %bb.0:
-; CM-NEXT: CF_END
-; CM-NEXT: PAD
- %result = call float @llvm.exp2.f32(float %in)
- ret float %result
-}
-
define float @v_exp2_f32_ninf(float %in) {
; SI-SDAG-LABEL: v_exp2_f32_ninf:
; SI-SDAG: ; %bb.0:
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.log2.ll b/llvm/test/CodeGen/AMDGPU/llvm.log2.ll
index 4ca612aa32e84..0854134be6f46 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.log2.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.log2.ll
@@ -2030,129 +2030,6 @@ define float @v_log2_f32_unsafe_math_attr(float %in) "unsafe-fp-math"="true" {
ret float %result
}
-define float @v_log2_f32_approx_fn_attr(float %in) "approx-func-fp-math"="true" {
-; SI-SDAG-LABEL: v_log2_f32_approx_fn_attr:
-; SI-SDAG: ; %bb.0:
-; SI-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; SI-SDAG-NEXT: s_mov_b32 s4, 0x800000
-; SI-SDAG-NEXT: v_cmp_gt_f32_e32 vcc, s4, v0
-; SI-SDAG-NEXT: v_cndmask_b32_e64 v2, 0, 32, vcc
-; SI-SDAG-NEXT: v_ldexp_f32_e32 v0, v0, v2
-; SI-SDAG-NEXT: v_log_f32_e32 v0, v0
-; SI-SDAG-NEXT: v_mov_b32_e32 v1, 0x42000000
-; SI-SDAG-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; SI-SDAG-NEXT: v_sub_f32_e32 v0, v0, v1
-; SI-SDAG-NEXT: s_setpc_b64 s[30:31]
-;
-; SI-GISEL-LABEL: v_log2_f32_approx_fn_attr:
-; SI-GISEL: ; %bb.0:
-; SI-GISEL-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; SI-GISEL-NEXT: v_mov_b32_e32 v1, 0x800000
-; SI-GISEL-NEXT: v_cmp_lt_f32_e32 vcc, v0, v1
-; SI-GISEL-NEXT: v_cndmask_b32_e64 v1, 0, 1, vcc
-; SI-GISEL-NEXT: v_lshlrev_b32_e32 v1, 5, v1
-; SI-GISEL-NEXT: v_ldexp_f32_e32 v0, v0, v1
-; SI-GISEL-NEXT: v_log_f32_e32 v0, v0
-; SI-GISEL-NEXT: v_mov_b32_e32 v1, 0x42000000
-; SI-GISEL-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; SI-GISEL-NEXT: v_sub_f32_e32 v0, v0, v1
-; SI-GISEL-NEXT: s_setpc_b64 s[30:31]
-;
-; VI-SDAG-LABEL: v_log2_f32_approx_fn_attr:
-; VI-SDAG: ; %bb.0:
-; VI-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; VI-SDAG-NEXT: s_mov_b32 s4, 0x800000
-; VI-SDAG-NEXT: v_cmp_gt_f32_e32 vcc, s4, v0
-; VI-SDAG-NEXT: v_cndmask_b32_e64 v2, 0, 32, vcc
-; VI-SDAG-NEXT: v_ldexp_f32 v0, v0, v2
-; VI-SDAG-NEXT: v_log_f32_e32 v0, v0
-; VI-SDAG-NEXT: v_mov_b32_e32 v1, 0x42000000
-; VI-SDAG-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; VI-SDAG-NEXT: v_sub_f32_e32 v0, v0, v1
-; VI-SDAG-NEXT: s_setpc_b64 s[30:31]
-;
-; VI-GISEL-LABEL: v_log2_f32_approx_fn_attr:
-; VI-GISEL: ; %bb.0:
-; VI-GISEL-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; VI-GISEL-NEXT: v_mov_b32_e32 v1, 0x800000
-; VI-GISEL-NEXT: v_cmp_lt_f32_e32 vcc, v0, v1
-; VI-GISEL-NEXT: v_cndmask_b32_e64 v1, 0, 1, vcc
-; VI-GISEL-NEXT: v_lshlrev_b32_e32 v1, 5, v1
-; VI-GISEL-NEXT: v_ldexp_f32 v0, v0, v1
-; VI-GISEL-NEXT: v_log_f32_e32 v0, v0
-; VI-GISEL-NEXT: v_mov_b32_e32 v1, 0x42000000
-; VI-GISEL-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; VI-GISEL-NEXT: v_sub_f32_e32 v0, v0, v1
-; VI-GISEL-NEXT: s_setpc_b64 s[30:31]
-;
-; GFX900-SDAG-LABEL: v_log2_f32_approx_fn_attr:
-; GFX900-SDAG: ; %bb.0:
-; GFX900-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX900-SDAG-NEXT: s_mov_b32 s4, 0x800000
-; GFX900-SDAG-NEXT: v_cmp_gt_f32_e32 vcc, s4, v0
-; GFX900-SDAG-NEXT: v_cndmask_b32_e64 v2, 0, 32, vcc
-; GFX900-SDAG-NEXT: v_ldexp_f32 v0, v0, v2
-; GFX900-SDAG-NEXT: v_log_f32_e32 v0, v0
-; GFX900-SDAG-NEXT: v_mov_b32_e32 v1, 0x42000000
-; GFX900-SDAG-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; GFX900-SDAG-NEXT: v_sub_f32_e32 v0, v0, v1
-; GFX900-SDAG-NEXT: s_setpc_b64 s[30:31]
-;
-; GFX900-GISEL-LABEL: v_log2_f32_approx_fn_attr:
-; GFX900-GISEL: ; %bb.0:
-; GFX900-GISEL-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX900-GISEL-NEXT: v_mov_b32_e32 v1, 0x800000
-; GFX900-GISEL-NEXT: v_cmp_lt_f32_e32 vcc, v0, v1
-; GFX900-GISEL-NEXT: v_cndmask_b32_e64 v1, 0, 1, vcc
-; GFX900-GISEL-NEXT: v_lshlrev_b32_e32 v1, 5, v1
-; GFX900-GISEL-NEXT: v_ldexp_f32 v0, v0, v1
-; GFX900-GISEL-NEXT: v_log_f32_e32 v0, v0
-; GFX900-GISEL-NEXT: v_mov_b32_e32 v1, 0x42000000
-; GFX900-GISEL-NEXT: v_cndmask_b32_e32 v1, 0, v1, vcc
-; GFX900-GISEL-NEXT: v_sub_f32_e32 v0, v0, v1
-; GFX900-GISEL-NEXT: s_setpc_b64 s[30:31]
-;
-; GFX1100-SDAG-LABEL: v_log2_f32_approx_fn_attr:
-; GFX1100-SDAG: ; %bb.0:
-; GFX1100-SDAG-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX1100-SDAG-NEXT: v_cmp_gt_f32_e32 vcc_lo, 0x800000, v0
-; GFX1100-SDAG-NEXT: v_cndmask_b32_e64 v2, 0, 32, vcc_lo
-; GFX1100-SDAG-NEXT: v_cndmask_b32_e64 v1, 0, 0x42000000, vcc_lo
-; GFX1100-SDAG-NEXT: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_1)
-; GFX1100-SDAG-NEXT: v_ldexp_f32 v0, v0, v2
-; GFX1100-SDAG-NEXT: v_log_f32_e32 v0, v0
-; GFX1100-SDAG-NEXT: s_waitcnt_depctr 0xfff
-; GFX1100-SDAG-NEXT: v_sub_f32_e32 v0, v0, v1
-; GFX1100-SDAG-NEXT: s_setpc_b64 s[30:31]
-;
-; GFX1100-GISEL-LABEL: v_log2_f32_approx_fn_attr:
-; GFX1100-GISEL: ; %bb.0:
-; GFX1100-GISEL-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX1100-GISEL-NEXT: v_cmp_gt_f32_e32 vcc_lo, 0x800000, v0
-; GFX1100-GISEL-NEXT: v_cndmask_b32_e64 v1, 0, 1, vcc_lo
-; GFX1100-GISEL-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
-; GFX1100-GISEL-NEXT: v_lshlrev_b32_e32 v1, 5, v1
-; GFX1100-GISEL-NEXT: v_ldexp_f32 v0, v0, v1
-; GFX1100-GISEL-NEXT: v_cndmask_b32_e64 v1, 0, 0x42000000, vcc_lo
-; GFX1100-GISEL-NEXT: s_delay_alu instid0(VALU_DEP_2)
-; GFX1100-GISEL-NEXT: v_log_f32_e32 v0, v0
-; GFX1100-GISEL-NEXT: s_waitcnt_depctr 0xfff
-; GFX1100-GISEL-NEXT: v_sub_f32_e32 v0, v0, v1
-; GFX1100-GISEL-NEXT: s_setpc_b64 s[30:31]
-;
-; R600-LABEL: v_log2_f32_approx_fn_attr:
-; R600: ; %bb.0:
-; R600-NEXT: CF_END
-; R600-NEXT: PAD
-;
-; CM-LABEL: v_log2_f32_approx_fn_attr:
-; CM: ; %bb.0:
-; CM-NEXT: CF_END
-; CM-NEXT: PAD
- %result = call float @llvm.log2.f32(float %in)
- ret float %result
-}
-
define float @v_log2_f32_ninf(float %in) {
; SI-SDAG-LABEL: v_log2_f32_ninf:
; SI-SDAG: ; %bb.0:
diff --git a/llvm/test/CodeGen/DirectX/ForwardHandleAccesses/cbuffer-access.ll b/llvm/test/CodeGen/DirectX/ForwardHandleAccesses/cbuffer-access.ll
index ca3c9427617a6..26b157f35a3e4 100644
--- a/llvm/test/CodeGen/DirectX/ForwardHandleAccesses/cbuffer-access.ll
+++ b/llvm/test/CodeGen/DirectX/ForwardHandleAccesses/cbuffer-access.ll
@@ -29,4 +29,4 @@ entry:
ret void
}
-attributes #0 = { mustprogress nofree noinline norecurse nosync nounwind willreturn memory(readwrite, argmem: write, inaccessiblemem: none) "approx-func-fp-math"="false" "frame-pointer"="all" "hlsl.numthreads"="1,1,1" "hlsl.shader"="compute" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
+attributes #0 = { mustprogress nofree noinline norecurse nosync nounwind willreturn memory(readwrite, argmem: write, inaccessiblemem: none) "frame-pointer"="all" "hlsl.numthreads"="1,1,1" "hlsl.shader"="compute" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
diff --git a/llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-cs.ll b/llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-cs.ll
index 3f2ec9a85b2a0..4bdb7ec50f6f4 100644
--- a/llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-cs.ll
+++ b/llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-cs.ll
@@ -30,5 +30,5 @@ entry:
ret i32 0
}
-attributes #0 = { convergent noinline norecurse optnone "approx-func-fp-math"="true" "hlsl.numthreads"="1,1,1" "hlsl.shader"="compute" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
-attributes #1 = { alwaysinline convergent mustprogress norecurse nounwind "approx-func-fp-math"="true" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
+attributes #0 = { convergent noinline norecurse optnone "hlsl.numthreads"="1,1,1" "hlsl.shader"="compute" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
+attributes #1 = { alwaysinline convergent mustprogress norecurse nounwind "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
diff --git a/llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-lib.ll b/llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-lib.ll
index c6e3cc9e96478..03756710adc3a 100644
--- a/llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-lib.ll
+++ b/llvm/test/CodeGen/DirectX/ShaderFlags/disable-opt-lib.ll
@@ -40,5 +40,5 @@ entry:
ret i32 0
}
-attributes #0 = { convergent mustprogress noinline norecurse noun...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/155966
More information about the llvm-commits
mailing list