[llvm] [AMDGPU] gfx1250 intrinscs codegen tests update. NFC (PR #155312)
via llvm-commits
llvm-commits at lists.llvm.org
Mon Aug 25 14:45:42 PDT 2025
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-amdgpu
Author: Stanislav Mekhanoshin (rampitec)
<details>
<summary>Changes</summary>
---
Patch is 28.83 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/155312.diff
4 Files Affected:
- (modified) llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.abs.ll (+186-13)
- (modified) llvm/test/CodeGen/AMDGPU/llvm.amdgcn.is.shared.ll (+48)
- (modified) llvm/test/CodeGen/AMDGPU/llvm.amdgcn.sendmsg.rtn.ll (+114)
- (modified) llvm/test/CodeGen/AMDGPU/llvm.amdgcn.wave.id.ll (+41-9)
``````````diff
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.abs.ll b/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.abs.ll
index 7916267c6eca2..800df89877036 100644
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.abs.ll
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.abs.ll
@@ -2,6 +2,7 @@
; RUN: llc -global-isel -mtriple=amdgcn -mcpu=tahiti -o - < %s | FileCheck %s --check-prefixes=GFX,GFX6
; RUN: llc -global-isel -mtriple=amdgcn -mcpu=fiji -o - < %s | FileCheck %s --check-prefixes=GFX,GFX8
; RUN: llc -global-isel -mtriple=amdgcn -mcpu=gfx1010 -o - < %s | FileCheck %s --check-prefixes=GFX,GFX10
+; RUN: llc -global-isel -mtriple=amdgcn -mcpu=gfx1250 -o - < %s | FileCheck %s --check-prefixes=GFX,GFX1250
declare i16 @llvm.abs.i16(i16, i1)
declare i32 @llvm.abs.i32(i32, i1)
@@ -13,11 +14,30 @@ declare <3 x i16> @llvm.abs.v3i16(<3 x i16>, i1)
declare <4 x i32> @llvm.abs.v4i32(<4 x i32>, i1)
define amdgpu_cs i16 @abs_sgpr_i16(i16 inreg %arg) {
-; GFX-LABEL: abs_sgpr_i16:
-; GFX: ; %bb.0:
-; GFX-NEXT: s_sext_i32_i16 s0, s0
-; GFX-NEXT: s_abs_i32 s0, s0
-; GFX-NEXT: ; return to shader part epilog
+; GFX6-LABEL: abs_sgpr_i16:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_sext_i32_i16 s0, s0
+; GFX6-NEXT: s_abs_i32 s0, s0
+; GFX6-NEXT: ; return to shader part epilog
+;
+; GFX8-LABEL: abs_sgpr_i16:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_sext_i32_i16 s0, s0
+; GFX8-NEXT: s_abs_i32 s0, s0
+; GFX8-NEXT: ; return to shader part epilog
+;
+; GFX10-LABEL: abs_sgpr_i16:
+; GFX10: ; %bb.0:
+; GFX10-NEXT: s_sext_i32_i16 s0, s0
+; GFX10-NEXT: s_abs_i32 s0, s0
+; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_sgpr_i16:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: s_sext_i32_i16 s0, s0
+; GFX1250-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
+; GFX1250-NEXT: s_abs_i32 s0, s0
+; GFX1250-NEXT: ; return to shader part epilog
%res = call i16 @llvm.abs.i16(i16 %arg, i1 false)
ret i16 %res
}
@@ -32,14 +52,42 @@ define amdgpu_cs i32 @abs_sgpr_i32(i32 inreg %arg) {
}
define amdgpu_cs i64 @abs_sgpr_i64(i64 inreg %arg) {
-; GFX-LABEL: abs_sgpr_i64:
-; GFX: ; %bb.0:
-; GFX-NEXT: s_ashr_i32 s2, s1, 31
-; GFX-NEXT: s_add_u32 s0, s0, s2
-; GFX-NEXT: s_mov_b32 s3, s2
-; GFX-NEXT: s_addc_u32 s1, s1, s2
-; GFX-NEXT: s_xor_b64 s[0:1], s[0:1], s[2:3]
-; GFX-NEXT: ; return to shader part epilog
+; GFX6-LABEL: abs_sgpr_i64:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_ashr_i32 s2, s1, 31
+; GFX6-NEXT: s_add_u32 s0, s0, s2
+; GFX6-NEXT: s_mov_b32 s3, s2
+; GFX6-NEXT: s_addc_u32 s1, s1, s2
+; GFX6-NEXT: s_xor_b64 s[0:1], s[0:1], s[2:3]
+; GFX6-NEXT: ; return to shader part epilog
+;
+; GFX8-LABEL: abs_sgpr_i64:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_ashr_i32 s2, s1, 31
+; GFX8-NEXT: s_add_u32 s0, s0, s2
+; GFX8-NEXT: s_mov_b32 s3, s2
+; GFX8-NEXT: s_addc_u32 s1, s1, s2
+; GFX8-NEXT: s_xor_b64 s[0:1], s[0:1], s[2:3]
+; GFX8-NEXT: ; return to shader part epilog
+;
+; GFX10-LABEL: abs_sgpr_i64:
+; GFX10: ; %bb.0:
+; GFX10-NEXT: s_ashr_i32 s2, s1, 31
+; GFX10-NEXT: s_add_u32 s0, s0, s2
+; GFX10-NEXT: s_mov_b32 s3, s2
+; GFX10-NEXT: s_addc_u32 s1, s1, s2
+; GFX10-NEXT: s_xor_b64 s[0:1], s[0:1], s[2:3]
+; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_sgpr_i64:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: s_ashr_i32 s2, s1, 31
+; GFX1250-NEXT: s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
+; GFX1250-NEXT: s_mov_b32 s3, s2
+; GFX1250-NEXT: s_add_nc_u64 s[0:1], s[0:1], s[2:3]
+; GFX1250-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
+; GFX1250-NEXT: s_xor_b64 s[0:1], s[0:1], s[2:3]
+; GFX1250-NEXT: ; return to shader part epilog
%res = call i64 @llvm.abs.i64(i64 %arg, i1 false)
ret i64 %res
}
@@ -78,6 +126,14 @@ define amdgpu_cs i16 @abs_vgpr_i16(i16 %arg) {
; GFX10-NEXT: v_max_i16 v0, v0, v1
; GFX10-NEXT: v_readfirstlane_b32 s0, v0
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_vgpr_i16:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: v_sub_nc_u16 v1, 0, v0
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; GFX1250-NEXT: v_max_i16 v0, v0, v1
+; GFX1250-NEXT: v_readfirstlane_b32 s0, v0
+; GFX1250-NEXT: ; return to shader part epilog
%res = call i16 @llvm.abs.i16(i16 %arg, i1 false)
ret i16 %res
}
@@ -103,6 +159,14 @@ define amdgpu_cs i32 @abs_vgpr_i32(i32 %arg) {
; GFX10-NEXT: v_max_i32_e32 v0, v0, v1
; GFX10-NEXT: v_readfirstlane_b32 s0, v0
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_vgpr_i32:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: v_sub_nc_u32_e32 v1, 0, v0
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; GFX1250-NEXT: v_max_i32_e32 v0, v0, v1
+; GFX1250-NEXT: v_readfirstlane_b32 s0, v0
+; GFX1250-NEXT: ; return to shader part epilog
%res = call i32 @llvm.abs.i32(i32 %arg, i1 false)
ret i32 %res
}
@@ -140,6 +204,20 @@ define amdgpu_cs i64 @abs_vgpr_i64(i64 %arg) {
; GFX10-NEXT: v_readfirstlane_b32 s0, v0
; GFX10-NEXT: v_readfirstlane_b32 s1, v1
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_vgpr_i64:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: v_ashrrev_i32_e32 v2, 31, v1
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; GFX1250-NEXT: v_mov_b32_e32 v3, v2
+; GFX1250-NEXT: v_add_nc_u64_e32 v[0:1], v[0:1], v[2:3]
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_2)
+; GFX1250-NEXT: v_xor_b32_e32 v0, v0, v2
+; GFX1250-NEXT: v_xor_b32_e32 v1, v1, v2
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
+; GFX1250-NEXT: v_readfirstlane_b32 s0, v0
+; GFX1250-NEXT: v_readfirstlane_b32 s1, v1
+; GFX1250-NEXT: ; return to shader part epilog
%res = call i64 @llvm.abs.i64(i64 %arg, i1 false)
ret i64 %res
}
@@ -192,6 +270,24 @@ define amdgpu_cs <4 x i32> @abs_vgpr_v4i32(<4 x i32> %arg) {
; GFX10-NEXT: v_readfirstlane_b32 s2, v2
; GFX10-NEXT: v_readfirstlane_b32 s3, v3
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_vgpr_v4i32:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: v_dual_sub_nc_u32 v4, 0, v0 :: v_dual_sub_nc_u32 v5, 0, v1
+; GFX1250-NEXT: v_dual_sub_nc_u32 v6, 0, v2 :: v_dual_sub_nc_u32 v7, 0, v3
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_3)
+; GFX1250-NEXT: v_max_i32_e32 v0, v0, v4
+; GFX1250-NEXT: v_max_i32_e32 v1, v1, v5
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_4)
+; GFX1250-NEXT: v_max_i32_e32 v2, v2, v6
+; GFX1250-NEXT: v_max_i32_e32 v3, v3, v7
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4)
+; GFX1250-NEXT: v_readfirstlane_b32 s0, v0
+; GFX1250-NEXT: v_readfirstlane_b32 s1, v1
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_4) | instskip(NEXT) | instid1(VALU_DEP_4)
+; GFX1250-NEXT: v_readfirstlane_b32 s2, v2
+; GFX1250-NEXT: v_readfirstlane_b32 s3, v3
+; GFX1250-NEXT: ; return to shader part epilog
%res = call <4 x i32> @llvm.abs.v4i32(<4 x i32> %arg, i1 false)
ret <4 x i32> %res
}
@@ -243,6 +339,21 @@ define amdgpu_cs <2 x i8> @abs_vgpr_v2i8(<2 x i8> %arg) {
; GFX10-NEXT: v_readfirstlane_b32 s0, v0
; GFX10-NEXT: v_readfirstlane_b32 s1, v1
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_vgpr_v2i8:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: v_bfe_i32 v0, v0, 0, 8
+; GFX1250-NEXT: v_bfe_i32 v1, v1, 0, 8
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
+; GFX1250-NEXT: v_sub_nc_u16 v2, 0, v0
+; GFX1250-NEXT: v_sub_nc_u16 v3, 0, v1
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
+; GFX1250-NEXT: v_max_i16 v0, v0, v2
+; GFX1250-NEXT: v_max_i16 v1, v1, v3
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
+; GFX1250-NEXT: v_readfirstlane_b32 s0, v0
+; GFX1250-NEXT: v_readfirstlane_b32 s1, v1
+; GFX1250-NEXT: ; return to shader part epilog
%res = call <2 x i8> @llvm.abs.v2i8(<2 x i8> %arg, i1 false)
ret <2 x i8> %res
}
@@ -307,6 +418,27 @@ define amdgpu_cs <3 x i8> @abs_vgpr_v3i8(<3 x i8> %arg) {
; GFX10-NEXT: v_readfirstlane_b32 s1, v1
; GFX10-NEXT: v_readfirstlane_b32 s2, v2
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_vgpr_v3i8:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: v_bfe_i32 v0, v0, 0, 8
+; GFX1250-NEXT: v_bfe_i32 v1, v1, 0, 8
+; GFX1250-NEXT: v_bfe_i32 v2, v2, 0, 8
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3)
+; GFX1250-NEXT: v_sub_nc_u16 v3, 0, v0
+; GFX1250-NEXT: v_sub_nc_u16 v4, 0, v1
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3)
+; GFX1250-NEXT: v_sub_nc_u16 v5, 0, v2
+; GFX1250-NEXT: v_max_i16 v0, v0, v3
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3)
+; GFX1250-NEXT: v_max_i16 v1, v1, v4
+; GFX1250-NEXT: v_max_i16 v2, v2, v5
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_3)
+; GFX1250-NEXT: v_readfirstlane_b32 s0, v0
+; GFX1250-NEXT: v_readfirstlane_b32 s1, v1
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_3)
+; GFX1250-NEXT: v_readfirstlane_b32 s2, v2
+; GFX1250-NEXT: ; return to shader part epilog
%res = call <3 x i8> @llvm.abs.v3i8(<3 x i8> %arg, i1 false)
ret <3 x i8> %res
}
@@ -341,6 +473,16 @@ define amdgpu_cs <2 x i16> @abs_sgpr_v2i16(<2 x i16> inreg %arg) {
; GFX10-NEXT: s_abs_i32 s0, s0
; GFX10-NEXT: s_pack_ll_b32_b16 s0, s1, s0
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_sgpr_v2i16:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: s_sext_i32_i16 s1, s0
+; GFX1250-NEXT: s_ashr_i32 s0, s0, 16
+; GFX1250-NEXT: s_abs_i32 s1, s1
+; GFX1250-NEXT: s_abs_i32 s0, s0
+; GFX1250-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
+; GFX1250-NEXT: s_pack_ll_b32_b16 s0, s1, s0
+; GFX1250-NEXT: ; return to shader part epilog
%res = call <2 x i16> @llvm.abs.v2i16(<2 x i16> %arg, i1 false)
ret <2 x i16> %res
}
@@ -375,6 +517,14 @@ define amdgpu_cs <2 x i16> @abs_vgpr_v2i16(<2 x i16> %arg) {
; GFX10-NEXT: v_pk_max_i16 v0, v0, v1
; GFX10-NEXT: v_readfirstlane_b32 s0, v0
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_vgpr_v2i16:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: v_pk_sub_i16 v1, 0, v0
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; GFX1250-NEXT: v_pk_max_i16 v0, v0, v1
+; GFX1250-NEXT: v_readfirstlane_b32 s0, v0
+; GFX1250-NEXT: ; return to shader part epilog
%res = call <2 x i16> @llvm.abs.v2i16(<2 x i16> %arg, i1 false)
ret <2 x i16> %res
}
@@ -416,6 +566,17 @@ define amdgpu_cs <3 x i16> @abs_sgpr_v3i16(<3 x i16> inreg %arg) {
; GFX10-NEXT: s_pack_ll_b32_b16 s0, s2, s0
; GFX10-NEXT: s_abs_i32 s1, s1
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_sgpr_v3i16:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: s_sext_i32_i16 s2, s0
+; GFX1250-NEXT: s_ashr_i32 s0, s0, 16
+; GFX1250-NEXT: s_abs_i32 s2, s2
+; GFX1250-NEXT: s_abs_i32 s0, s0
+; GFX1250-NEXT: s_sext_i32_i16 s1, s1
+; GFX1250-NEXT: s_pack_ll_b32_b16 s0, s2, s0
+; GFX1250-NEXT: s_abs_i32 s1, s1
+; GFX1250-NEXT: ; return to shader part epilog
%res = call <3 x i16> @llvm.abs.v3i16(<3 x i16> %arg, i1 false)
ret <3 x i16> %res
}
@@ -460,6 +621,18 @@ define amdgpu_cs <3 x i16> @abs_vgpr_v3i16(<3 x i16> %arg) {
; GFX10-NEXT: v_readfirstlane_b32 s0, v0
; GFX10-NEXT: v_readfirstlane_b32 s1, v1
; GFX10-NEXT: ; return to shader part epilog
+;
+; GFX1250-LABEL: abs_vgpr_v3i16:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: v_pk_sub_i16 v2, 0, v0
+; GFX1250-NEXT: v_sub_nc_u16 v3, 0, v1
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
+; GFX1250-NEXT: v_pk_max_i16 v0, v0, v2
+; GFX1250-NEXT: v_max_i16 v1, v1, v3
+; GFX1250-NEXT: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
+; GFX1250-NEXT: v_readfirstlane_b32 s0, v0
+; GFX1250-NEXT: v_readfirstlane_b32 s1, v1
+; GFX1250-NEXT: ; return to shader part epilog
%res = call <3 x i16> @llvm.abs.v3i16(<3 x i16> %arg, i1 false)
ret <3 x i16> %res
}
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.is.shared.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.is.shared.ll
index c364c391559ea..f1dcc93172fb1 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.is.shared.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.is.shared.ll
@@ -2,10 +2,12 @@
; RUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=tahiti < %s | FileCheck -check-prefixes=SI,SI-SDAG %s
; RUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=hawaii < %s | FileCheck -check-prefixes=CI,CI-SDAG %s
; RUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 < %s | FileCheck -check-prefixes=GFX9,GFX9-SDAG %s
+; RUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 < %s | FileCheck -check-prefixes=GFX1250,GFX1250-SDAG %s
; RUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=hawaii < %s | FileCheck -check-prefixes=CI,CI-GISEL %s
; RUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 < %s | FileCheck -check-prefixes=GFX9,GFX9-GISEL %s
; RUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1010 < %s | FileCheck -check-prefixes=GFX10,GFX10-GISEL %s
; RUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1100 < %s | FileCheck -check-prefixes=GFX11,GFX11-GISEL %s
+; RUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 < %s | FileCheck -check-prefixes=GFX1250,GFX1250-GISEL %s
define amdgpu_kernel void @is_local_vgpr(ptr addrspace(1) %ptr.ptr) {
; CIT-LABEL: is_local_vgpr:
@@ -90,6 +92,21 @@ define amdgpu_kernel void @is_local_vgpr(ptr addrspace(1) %ptr.ptr) {
; GFX9-NEXT: global_store_dword v[0:1], v0, off
; GFX9-NEXT: s_endpgm
;
+; GFX1250-LABEL: is_local_vgpr:
+; GFX1250: ; %bb.0:
+; GFX1250-NEXT: s_load_b64 s[0:1], s[4:5], 0x0
+; GFX1250-NEXT: v_and_b32_e32 v0, 0x3ff, v0
+; GFX1250-NEXT: s_wait_kmcnt 0x0
+; GFX1250-NEXT: global_load_b64 v[0:1], v0, s[0:1] scale_offset scope:SCOPE_SYS
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_mov_b64 s[0:1], src_shared_base
+; GFX1250-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
+; GFX1250-NEXT: v_cmp_eq_u32_e32 vcc_lo, s1, v1
+; GFX1250-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc_lo
+; GFX1250-NEXT: global_store_b32 v[0:1], v0, off
+; GFX1250-NEXT: s_endpgm
+;
; CI-GISEL-LABEL: is_local_vgpr:
; CI-GISEL: ; %bb.0:
; CI-GISEL-NEXT: s_load_dwordx2 s[0:1], s[8:9], 0x0
@@ -237,6 +254,23 @@ define amdgpu_kernel void @is_local_sgpr(ptr %ptr) {
; GFX9-SDAG-NEXT: .LBB1_2: ; %bb1
; GFX9-SDAG-NEXT: s_endpgm
;
+; GFX1250-SDAG-LABEL: is_local_sgpr:
+; GFX1250-SDAG: ; %bb.0:
+; GFX1250-SDAG-NEXT: s_load_b32 s2, s[4:5], 0x4
+; GFX1250-SDAG-NEXT: s_mov_b64 s[0:1], src_shared_base
+; GFX1250-SDAG-NEXT: s_wait_kmcnt 0x0
+; GFX1250-SDAG-NEXT: s_cmp_eq_u32 s2, s1
+; GFX1250-SDAG-NEXT: s_cselect_b32 s0, -1, 0
+; GFX1250-SDAG-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
+; GFX1250-SDAG-NEXT: s_and_not1_b32 vcc_lo, exec_lo, s0
+; GFX1250-SDAG-NEXT: s_cbranch_vccnz .LBB1_2
+; GFX1250-SDAG-NEXT: ; %bb.1: ; %bb0
+; GFX1250-SDAG-NEXT: v_mov_b32_e32 v0, 0
+; GFX1250-SDAG-NEXT: global_store_b32 v[0:1], v0, off scope:SCOPE_SYS
+; GFX1250-SDAG-NEXT: s_wait_storecnt 0x0
+; GFX1250-SDAG-NEXT: .LBB1_2: ; %bb1
+; GFX1250-SDAG-NEXT: s_endpgm
+;
; CI-GISEL-LABEL: is_local_sgpr:
; CI-GISEL: ; %bb.0:
; CI-GISEL-NEXT: s_load_dwordx2 s[0:1], s[8:9], 0x0
@@ -296,6 +330,20 @@ define amdgpu_kernel void @is_local_sgpr(ptr %ptr) {
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: .LBB1_2: ; %bb1
; GFX11-NEXT: s_endpgm
+;
+; GFX1250-GISEL-LABEL: is_local_sgpr:
+; GFX1250-GISEL: ; %bb.0:
+; GFX1250-GISEL-NEXT: s_load_b64 s[0:1], s[4:5], 0x0
+; GFX1250-GISEL-NEXT: s_mov_b64 s[2:3], src_shared_base
+; GFX1250-GISEL-NEXT: s_wait_kmcnt 0x0
+; GFX1250-GISEL-NEXT: s_cmp_lg_u32 s1, s3
+; GFX1250-GISEL-NEXT: s_cbranch_scc1 .LBB1_2
+; GFX1250-GISEL-NEXT: ; %bb.1: ; %bb0
+; GFX1250-GISEL-NEXT: v_mov_b32_e32 v0, 0
+; GFX1250-GISEL-NEXT: global_store_b32 v[0:1], v0, off scope:SCOPE_SYS
+; GFX1250-GISEL-NEXT: s_wait_storecnt 0x0
+; GFX1250-GISEL-NEXT: .LBB1_2: ; %bb1
+; GFX1250-GISEL-NEXT: s_endpgm
%val = call i1 @llvm.amdgcn.is.shared(ptr %ptr)
br i1 %val, label %bb0, label %bb1
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.sendmsg.rtn.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.sendmsg.rtn.ll
index e4a87e3250ca0..d7f057fa427ef 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.sendmsg.rtn.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.sendmsg.rtn.ll
@@ -1,6 +1,8 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -global-isel=0 -mtriple=amdgcn -mcpu=gfx1100 < %s | FileCheck --check-prefixes=GFX11,GFX11-SDAG %s
; RUN: llc -global-isel=1 -mtriple=amdgcn -mcpu=gfx1100 < %s | FileCheck --check-prefixes=GFX11,GFX11-GISEL %s
+; RUN: llc -global-isel=0 -mtriple=amdgcn -mcpu=gfx1250 < %s | FileCheck --check-prefixes=GFX1250,GFX1250-SDAG %s
+; RUN: llc -global-isel=1 -mtriple=amdgcn -mcpu=gfx1250 < %s | FileCheck --check-prefixes=GFX1250,GFX1250-GISEL %s
define amdgpu_kernel void @test_get_doorbell(ptr addrspace(1) %out) {
; GFX11-SDAG-LABEL: test_get_doorbell:
@@ -20,6 +22,24 @@ define amdgpu_kernel void @test_get_doorbell(ptr addrspace(1) %out) {
; GFX11-GISEL-NEXT: v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v0, s2
; GFX11-GISEL-NEXT: global_store_b32 v1, v0, s[0:1]
; GFX11-GISEL-NEXT: s_endpgm
+;
+; GFX1250-SDAG-LABEL: test_get_doorbell:
+; GFX1250-SDAG: ; %bb.0:
+; GFX1250-SDAG-NEXT: s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1250-SDAG-NEXT: s_sendmsg_rtn_b32 s2, sendmsg(MSG_RTN_GET_DOORBELL)
+; GFX1250-SDAG-NEXT: s_wait_kmcnt 0x0
+; GFX1250-SDAG-NEXT: v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s2
+; GFX1250-SDAG-NEXT: global_store_b32 v0, v1, s[0:1]
+; GFX1250-SDAG-NEXT: s_endpgm
+;
+; GFX1250-GISEL-LABEL: test_get_doorbell:
+; GFX1250-GISEL: ; %bb.0:
+; GFX1250-GISEL-NEXT: s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1250-GISEL-NEXT: s_sendmsg_rtn_b32 s2, sendmsg(MSG_RTN_GET_DOORBELL)
+; GFX1250-GISEL-NEXT: s_wait_kmcnt 0x0
+; GFX1250-GISEL-NEXT: v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v0, s2
+; GFX1250-GISEL-NEXT: global_store_b32 v1, v0, s[0:1]
+; GFX1250-GISEL-NEXT: s_endpgm
%ret = call i32 @llvm.amdgcn.s.sendmsg.rtn.i32(i32 128)
store i32 %ret, ptr addrspace(1) %out
ret void
@@ -43,6 +63,24 @@ define amdgpu_kernel void @test_get_ddid(ptr addrspace(1) %out) {
; GFX11-GISEL-NEXT: v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v0, s2
; GFX11-GISEL-NEXT: global_store_b32 v1, v0, s[0:1]
; GFX11-GISEL-NEXT: s_endpgm
+;
+; GFX1250-SDAG-LABEL: test_get_ddid:
+; GFX1250-SDAG: ; %bb.0:
+; GFX1250-SDAG-NEXT: s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1250-SDAG-NEXT: s_sendmsg_rtn_b32 s2, sendmsg(MSG_RTN_GET_DDID)
+; GFX1250-SDAG-NEXT: s_wait_kmcnt 0x0
+; GFX1250-SDAG-NEXT: v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s2
+; GFX1250-SDAG-NEXT: global_store_b32 v0, v1, s[0:1]
+; GFX1250-SDAG-NEXT: s_endpgm
+;
+; GFX1250-GISEL-LABEL: test_get_ddid:
+; GFX1250-GISEL: ; %bb.0:
+; GFX1250-GISEL-NEXT: s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1250-GISEL-NEXT: s_sendmsg_rtn_b32 s2, sendmsg(MSG_RTN_GET_DDID)
+; GFX1250-GISEL-NEXT: s_wait_kmcnt 0x0
+; GFX1250-GISEL-NEXT: v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v0, s2
+; GFX1250-GISEL-NEXT: global_store_b32 v1, v0, s[0:1]
+; GFX1250-GISEL-NEXT: s_endpgm
%ret = call i32 @llvm.amdgcn.s.sendmsg.rtn.i32(i32 129)
store i32 %ret, ptr addrspace(1) %out
ret void
@@ -58,6 +96,16 @@ define amdgpu_kernel void @test_get_tma(ptr addrspace(1) %out) {
; GFX11-NEXT: v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/155312
More information about the llvm-commits
mailing list