[llvm] [AMDGPU] Improve VGPR lowering test around FMA[AK|MK]. NFC (PR #170633)
via llvm-commits
llvm-commits at lists.llvm.org
Thu Dec 4 01:44:42 PST 2025
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-amdgpu
Author: Stanislav Mekhanoshin (rampitec)
<details>
<summary>Changes</summary>
Add asm comments checks for readability.
---
Full diff: https://github.com/llvm/llvm-project/pull/170633.diff
1 Files Affected:
- (modified) llvm/test/CodeGen/AMDGPU/vgpr-lowering-gfx1250.mir (+25)
``````````diff
diff --git a/llvm/test/CodeGen/AMDGPU/vgpr-lowering-gfx1250.mir b/llvm/test/CodeGen/AMDGPU/vgpr-lowering-gfx1250.mir
index 21f5515b7fb91..10c9d1eb9f6d7 100644
--- a/llvm/test/CodeGen/AMDGPU/vgpr-lowering-gfx1250.mir
+++ b/llvm/test/CodeGen/AMDGPU/vgpr-lowering-gfx1250.mir
@@ -260,50 +260,62 @@ body: |
$vgpr255, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32_gfx1250 undef $vgpr1, undef $vgpr1, undef $vgpr0, undef $vgpr0, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 64
+ ; ASM-SAME: ; msbs: dst=1 src0=0 src1=0 src2=0
; GCN-NEXT: v_dual_sub_f32 v244 /*v500*/, v1, v2 :: v_dual_mul_f32 v0 /*v256*/, v3, v4
$vgpr500, $vgpr256 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32_gfx1250 undef $vgpr1, undef $vgpr2, undef $vgpr3, undef $vgpr4, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 0x4041
+ ; ASM-SAME: ; msbs: dst=1 src0=1 src1=0 src2=0
; GCN-NEXT: v_dual_sub_f32 v244 /*v500*/, s1, v2 :: v_dual_mul_f32 v0 /*v256*/, v44 /*v300*/, v4
$vgpr500, $vgpr256 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32_gfx1250 undef $sgpr1, undef $vgpr2, undef $vgpr300, undef $vgpr4, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 0x4104
+ ; ASM-SAME: ; msbs: dst=0 src0=0 src1=1 src2=0
; GCN-NEXT: v_dual_sub_f32 v255, v1, v44 /*v300*/ :: v_dual_mul_f32 v6, v0, v1 /*v257*/
$vgpr255, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32_gfx1250 undef $vgpr1, undef $vgpr300, undef $vgpr0, $vgpr257, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 0x401
+ ; ASM-SAME: ; msbs: dst=0 src0=1 src1=0 src2=0
; GCN-NEXT: v_dual_sub_f32 v255, 0, v1 :: v_dual_mul_f32 v6, v44 /*v300*/, v3
$vgpr255, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32_gfx1250 0, undef $vgpr1, undef $vgpr300, undef $vgpr3, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 0x140
+ ; ASM-SAME: ; msbs: dst=1 src0=0 src1=0 src2=0
; GCN-NEXT: v_dual_fmamk_f32 v243 /*v499*/, v0, 0xa, v3 :: v_dual_fmac_f32 v0 /*v256*/, v1, v1
$vgpr499, $vgpr256 = V_DUAL_FMAMK_F32_X_FMAC_F32_e32_gfx1250 undef $vgpr0, 10, undef $vgpr3, undef $vgpr1, undef $vgpr1, $vgpr256, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 0x4005
+ ; ASM-SAME: ; msbs: dst=0 src0=1 src1=1 src2=0
; GCN-NEXT: v_dual_mov_b32 v2, v3 /*v259*/ :: v_dual_add_f32 v3, v1 /*v257*/, v2 /*v258*/
$vgpr2, $vgpr3 = V_DUAL_MOV_B32_e32_X_ADD_F32_e32_gfx1250 undef $vgpr259, undef $vgpr257, undef $vgpr258, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x554
+ ; ASM-SAME: ; msbs: dst=1 src0=0 src1=1 src2=1
; GCN-NEXT: v_dual_fmamk_f32 v244 /*v500*/, v0, 0xa, v44 /*v300*/ :: v_dual_fmac_f32 v3 /*v259*/, v1, v1 /*v257*/
$vgpr500, $vgpr259 = V_DUAL_FMAMK_F32_X_FMAC_F32_e32_gfx1250 undef $vgpr0, 10, undef $vgpr300, undef $vgpr1, undef $vgpr257, $vgpr259, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 0x5410
+ ; ASM-SAME: ; msbs: dst=0 src0=0 src1=0 src2=1
; GCN-NEXT: v_dual_fma_f32 v0, v6, v6, v44 /*v300*/ :: v_dual_fma_f32 v1, v4, v5, v45 /*v301*/
$vgpr0, $vgpr1 = V_DUAL_FMA_F32_e64_X_FMA_F32_e64_e96_gfx1250 0, undef $vgpr6, 0, undef $vgpr6, 0, undef $vgpr300, 0, undef $vgpr4, 0, undef $vgpr5, 0, undef $vgpr301, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 0x1000
+ ; ASM-SAME: ; msbs: dst=0 src0=0 src1=0 src2=0
; GCN-NEXT: v_dual_fmac_f32 v2, v6, v6 :: v_dual_fma_f32 v3, v4, v5, v3
$vgpr2, $vgpr3 = V_DUAL_FMAC_F32_e32_X_FMA_F32_e64_e96_gfx1250 0, undef $vgpr6, 0, undef $vgpr6, undef $vgpr2, 0, undef $vgpr4, 0, undef $vgpr5, 0, $vgpr3, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 64
+ ; ASM-SAME: ; msbs: dst=1 src0=0 src1=0 src2=0
; GCN-NEXT: v_dual_fma_f32 v244 /*v500*/, v6, v7, v8 :: v_dual_add_f32 v3 /*v259*/, v4, v5
$vgpr500, $vgpr259 = V_DUAL_FMA_F32_e64_X_ADD_F32_e32_e96_gfx1250 0, undef $vgpr6, 0, undef $vgpr7, 0, undef $vgpr8, 0, undef $vgpr4, 0, undef $vgpr5, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 0x40ae
+ ; ASM-SAME: ; msbs: dst=2 src0=2 src1=3 src2=2
; GCN-NEXT: v_dual_fmac_f32 v2 /*v514*/, v6 /*v518*/, v8 /*v776*/ :: v_dual_fma_f32 v3 /*v515*/, v4 /*v516*/, v7 /*v775*/, v3 /*v515*/
$vgpr514, $vgpr515 = V_DUAL_FMAC_F32_e32_X_FMA_F32_e64_e96_gfx1250 0, undef $vgpr518, 0, undef $vgpr776, undef $vgpr514, 0, undef $vgpr516, 0, undef $vgpr775, 0, $vgpr515, implicit $mode, implicit $exec
; GCN-NEXT: s_set_vgpr_msb 0xae54
+ ; ASM-SAME: ; msbs: dst=1 src0=0 src1=1 src2=1
; GCN-NEXT: v_dual_fmac_f32 v7 /*v263*/, v1, v1 /*v257*/ :: v_dual_fmamk_f32 v244 /*v500*/, v0, 0xa, v44 /*v300*/
$vgpr263, $vgpr500 = V_DUAL_FMAC_F32_e32_X_FMAMK_F32_gfx1250 undef $vgpr1, undef $vgpr257, $vgpr263, undef $vgpr0, 10, undef $vgpr300, implicit $mode, implicit $exec
@@ -321,54 +333,67 @@ body: |
; ASM: %bb.0:
; GCN-NEXT: s_set_vgpr_msb 0x45
+ ; ASM-SAME: ; msbs: dst=1 src0=1 src1=1 src2=0
; GCN-NEXT: v_fmaak_f32 v0 /*v256*/, v1 /*v257*/, v2 /*v258*/, 0x1
$vgpr256 = V_FMAAK_F32 undef $vgpr257, undef $vgpr258, 1, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x4505
+ ; ASM-SAME: ; msbs: dst=0 src0=1 src1=1 src2=0
; GCN-NEXT: v_fmaak_f32 v0, v1 /*v257*/, v2 /*v258*/, 0x1
$vgpr0 = V_FMAAK_F32 undef $vgpr257, undef $vgpr258, 1, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x541
+ ; ASM-SAME: ; msbs: dst=1 src0=1 src1=0 src2=0
; GCN-NEXT: v_fmaak_f32 v0 /*v256*/, v1 /*v257*/, v2, 0x1
$vgpr256 = V_FMAAK_F32 undef $vgpr257, undef $vgpr2, 1, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x4144
+ ; ASM-SAME: ; msbs: dst=1 src0=0 src1=1 src2=0
; GCN-NEXT: v_fmaak_f32 v0 /*v256*/, v1, v2 /*v258*/, 0x1
$vgpr256 = V_FMAAK_F32 undef $vgpr1, undef $vgpr258, 1, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x4451
+ ; ASM-SAME: ; msbs: dst=1 src0=1 src1=0 src2=1
; GCN-NEXT: v_fmamk_f32 v0 /*v256*/, v1 /*v257*/, 0x1, v2 /*v258*/
$vgpr256 = V_FMAMK_F32 undef $vgpr257, 1, undef $vgpr258, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x5111
+ ; ASM-SAME: ; msbs: dst=0 src0=1 src1=0 src2=1
; GCN-NEXT: v_fmamk_f32 v0, v1 /*v257*/, 0x1, v2 /*v258*/
$vgpr0 = V_FMAMK_F32 undef $vgpr257, 1, undef $vgpr258, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x1141
+ ; ASM-SAME: ; msbs: dst=1 src0=1 src1=0 src2=0
; GCN-NEXT: v_fmamk_f32 v0 /*v256*/, v1 /*v257*/, 0x1, v2
$vgpr256 = V_FMAMK_F32 undef $vgpr257, 1, undef $vgpr2, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x4150
+ ; ASM-SAME: ; msbs: dst=1 src0=0 src1=0 src2=1
; GCN-NEXT: v_fmamk_f32 v0 /*v256*/, v1, 0x1, v2 /*v258*/
$vgpr256 = V_FMAMK_F32 undef $vgpr1, 1, undef $vgpr258, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x5051
+ ; ASM-SAME: ; msbs: dst=1 src0=1 src1=0 src2=1
; GCN-NEXT: v_fmamk_f64 v[4:5] /*v[260:261]*/, v[100:101] /*v[356:357]*/, 0x1, v[2:3] /*v[258:259]*/
$vgpr260_vgpr261 = V_FMAMK_F64 undef $vgpr356_vgpr357, 1, undef $vgpr258_vgpr259, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x5101
+ ; ASM-SAME: ; msbs: dst=0 src0=1 src1=0 src2=0
; GCN-NEXT: v_fmamk_f64 v[0:1], v[100:101] /*v[356:357]*/, 0x1, v[2:3]
$vgpr0_vgpr1 = V_FMAMK_F64 undef $vgpr356_vgpr357, 1, undef $vgpr2_vgpr3, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x110
+ ; ASM-SAME: ; msbs: dst=0 src0=0 src1=0 src2=1
; GCN-NEXT: v_fmamk_f64 v[0:1], v[2:3], 0x1, v[100:101] /*v[356:357]*/
$vgpr0_vgpr1 = V_FMAMK_F64 undef $vgpr2_vgpr3, 1, undef $vgpr356_vgpr357, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x1040
+ ; ASM-SAME: ; msbs: dst=1 src0=0 src1=0 src2=0
; GCN-NEXT: v_fmamk_f64 v[0:1] /*v[256:257]*/, v[2:3], 0x1, v[4:5]
$vgpr256_vgpr257 = V_FMAMK_F64 undef $vgpr2_vgpr3, 1, undef $vgpr4_vgpr5, implicit $exec, implicit $mode
; GCN-NEXT: s_set_vgpr_msb 0x4000
+ ; ASM-SAME: ; msbs: dst=0 src0=0 src1=0 src2=0
; GCN-NEXT: v_fmamk_f16 v26, v56, 0x1, v58
$vgpr26 = V_FMAMK_F16_fake16 undef $vgpr56, 1, undef $vgpr58, implicit $exec, implicit $mode
``````````
</details>
https://github.com/llvm/llvm-project/pull/170633
More information about the llvm-commits
mailing list