[llvm] AMDGPU/GlobalISel: Add regbanklegalize rules for uniform global loads (PR #145909)
Matt Arsenault via llvm-commits
llvm-commits at lists.llvm.org
Thu Jun 26 23:13:20 PDT 2025
================
@@ -0,0 +1,95 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=gfx1010 -verify-machineinstrs < %s | FileCheck %s
+
+define amdgpu_ps void @uniform_load_32(ptr addrspace(1) inreg %ptr0, ptr addrspace(1) inreg %ptr1, ptr addrspace(1) inreg %ptr2) {
+; CHECK-LABEL: uniform_load_32:
+; CHECK: ; %bb.0:
+; CHECK-NEXT: v_mov_b32_e32 v0, 0
+; CHECK-NEXT: global_load_dword v1, v0, s[0:1] glc dlc
+; CHECK-NEXT: s_waitcnt vmcnt(0)
+; CHECK-NEXT: global_load_dword v2, v0, s[2:3] glc dlc
+; CHECK-NEXT: s_waitcnt vmcnt(0)
+; CHECK-NEXT: v_readfirstlane_b32 s0, v1
+; CHECK-NEXT: v_readfirstlane_b32 s1, v2
+; CHECK-NEXT: s_add_i32 s0, s0, s1
+; CHECK-NEXT: v_mov_b32_e32 v1, s0
+; CHECK-NEXT: global_store_dword v0, v1, s[4:5]
+; CHECK-NEXT: s_endpgm
+ %load0 = load volatile i32, ptr addrspace(1) %ptr0
+ %load1 = load volatile i32, ptr addrspace(1) %ptr1
+ %sum = add i32 %load0, %load1
+ store i32 %sum, ptr addrspace(1) %ptr2
+ ret void
+}
+
+define amdgpu_ps void @uniform_load_64(ptr addrspace(1) inreg %ptr0, ptr addrspace(1) inreg %ptr1) {
+; CHECK-LABEL: uniform_load_64:
+; CHECK: ; %bb.0:
+; CHECK-NEXT: v_mov_b32_e32 v2, 0
+; CHECK-NEXT: global_load_dwordx2 v[0:1], v2, s[0:1] glc dlc
+; CHECK-NEXT: s_waitcnt vmcnt(0)
+; CHECK-NEXT: v_readfirstlane_b32 s0, v0
+; CHECK-NEXT: v_readfirstlane_b32 s1, v1
+; CHECK-NEXT: s_add_i32 s0, s0, s1
+; CHECK-NEXT: v_mov_b32_e32 v0, s0
+; CHECK-NEXT: global_store_dword v2, v0, s[2:3]
+; CHECK-NEXT: s_endpgm
+ %load = load volatile <2 x i32>, ptr addrspace(1) %ptr0
+ %elt0 = extractelement <2 x i32> %load, i32 0
+ %elt1 = extractelement <2 x i32> %load, i32 1
+ %sum = add i32 %elt0, %elt1
+ store i32 %sum, ptr addrspace(1) %ptr1
+ ret void
+}
+
+define amdgpu_ps void @uniform_load_96(ptr addrspace(1) inreg %ptr0, ptr addrspace(1) inreg %ptr1) {
----------------
arsenm wrote:
```suggestion
define amdgpu_ps void @uniform_load_v3i32(ptr addrspace(1) inreg %ptr0, ptr addrspace(1) inreg %ptr1) {
```
https://github.com/llvm/llvm-project/pull/145909
More information about the llvm-commits
mailing list