[llvm] [AMDGPU][True16][CodeGen] buildbot fix for uaddsat/usubsat true16 selection (PR #128784)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-amdgpu

Author: Brox Chen (broxigarchen)

<details>
<summary>Changes</summary>



---
Full diff: https://github.com/llvm/llvm-project/pull/128784.diff


1 Files Affected:

- (modified) llvm/test/CodeGen/AMDGPU/fmed3.ll (+4-7) 


``````````diff
diff --git a/llvm/test/CodeGen/AMDGPU/fmed3.ll b/llvm/test/CodeGen/AMDGPU/fmed3.ll
index 57c692621c121..c70b921c97999 100644
--- a/llvm/test/CodeGen/AMDGPU/fmed3.ll
+++ b/llvm/test/CodeGen/AMDGPU/fmed3.ll
@@ -7582,7 +7582,7 @@ define amdgpu_kernel void @v_test_nnan_input_fmed3_r_i_i_f16(ptr addrspace(1) %o
 ; GFX11-GISEL-TRUE16-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_4) | instid1(VALU_DEP_1)
 ; GFX11-GISEL-TRUE16-NEXT:    v_lshlrev_b32_e32 v1, 1, v0
 ; GFX11-GISEL-TRUE16-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX11-GISEL-TRUE16-NEXT:    global_load_u16 v0, v1, s[2:3]
+; GFX11-GISEL-TRUE16-NEXT:    global_load_d16_b16 v0, v1, s[2:3]
 ; GFX11-GISEL-TRUE16-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-GISEL-TRUE16-NEXT:    v_add_f16_e32 v0.l, 1.0, v0.l
 ; GFX11-GISEL-TRUE16-NEXT:    v_med3_f16 v0.l, v0.l, 2.0, 4.0
@@ -7837,16 +7837,13 @@ define amdgpu_kernel void @v_nnan_inputs_med3_f16_pat0(ptr addrspace(1) %out, pt
 ; GFX11-GISEL-TRUE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-GISEL-TRUE16-NEXT:    v_lshlrev_b32_e32 v2, 1, v0
 ; GFX11-GISEL-TRUE16-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX11-GISEL-TRUE16-NEXT:    global_load_u16 v0, v2, s[2:3] glc dlc
+; GFX11-GISEL-TRUE16-NEXT:    global_load_d16_b16 v0, v2, s[2:3] glc dlc
 ; GFX11-GISEL-TRUE16-NEXT:    s_waitcnt vmcnt(0)
-; GFX11-GISEL-TRUE16-NEXT:    global_load_u16 v1, v2, s[4:5] glc dlc
+; GFX11-GISEL-TRUE16-NEXT:    global_load_d16_hi_b16 v0, v2, s[4:5] glc dlc
 ; GFX11-GISEL-TRUE16-NEXT:    s_waitcnt vmcnt(0)
-; GFX11-GISEL-TRUE16-NEXT:    global_load_u16 v3, v2, s[6:7] glc dlc
+; GFX11-GISEL-TRUE16-NEXT:    global_load_d16_b16 v1, v2, s[6:7] glc dlc
 ; GFX11-GISEL-TRUE16-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-GISEL-TRUE16-NEXT:    v_add_f16_e32 v0.l, 1.0, v0.l
-; GFX11-GISEL-TRUE16-NEXT:    v_mov_b16_e32 v0.h, v1.l
-; GFX11-GISEL-TRUE16-NEXT:    v_mov_b16_e32 v1.l, v3.l
-; GFX11-GISEL-TRUE16-NEXT:    s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
 ; GFX11-GISEL-TRUE16-NEXT:    v_add_f16_e32 v0.h, 2.0, v0.h
 ; GFX11-GISEL-TRUE16-NEXT:    v_add_f16_e32 v1.l, 4.0, v1.l
 ; GFX11-GISEL-TRUE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)

``````````

</details>


https://github.com/llvm/llvm-project/pull/128784