[llvm-branch-commits] [llvm] [AMDGPU] Add tests for workgroup/workitem intrinsic optimizations (PR #146053)
Pierre van Houtryve via llvm-branch-commits
llvm-branch-commits at lists.llvm.org
Fri Jun 27 03:46:00 PDT 2025
https://github.com/Pierre-vh created https://github.com/llvm/llvm-project/pull/146053
None
>From 3f62ab3beb30abbf8c8c32dd79c0133f7ca122e0 Mon Sep 17 00:00:00 2001
From: pvanhout <pierre.vanhoutryve at amd.com>
Date: Thu, 26 Jun 2025 13:08:31 +0200
Subject: [PATCH] [AMDGPU] Add tests for workgroup/workitem intrinsic
optimizations
---
.../AMDGPU/workitems-intrinsics-opts.ll | 553 ++++++++++++++++++
1 file changed, 553 insertions(+)
create mode 100644 llvm/test/CodeGen/AMDGPU/workitems-intrinsics-opts.ll
diff --git a/llvm/test/CodeGen/AMDGPU/workitems-intrinsics-opts.ll b/llvm/test/CodeGen/AMDGPU/workitems-intrinsics-opts.ll
new file mode 100644
index 0000000000000..14120680216fc
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/workitems-intrinsics-opts.ll
@@ -0,0 +1,553 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -O3 -mtriple=amdgcn -mcpu=fiji %s -o - | FileCheck %s --check-prefixes=GFX8,DAGISEL-GFX9
+; RUN: llc -O3 -mtriple=amdgcn -mcpu=gfx942 %s -o - | FileCheck %s --check-prefixes=GFX942,DAGISEL-GFX942
+; RUN: llc -O3 -mtriple=amdgcn -mcpu=gfx1200 %s -o - | FileCheck %s --check-prefixes=GFX12,DAGISEL-GFX12
+
+; RUN: llc -O3 -global-isel -mtriple=amdgcn -mcpu=fiji %s -o - | FileCheck %s --check-prefixes=GFX8,GISEL-GFX8
+; RUN: llc -O3 -global-isel -mtriple=amdgcn -mcpu=gfx942 %s -o - | FileCheck %s --check-prefixes=GFX942,GISEL-GFX942
+; RUN: llc -O3 -global-isel -mtriple=amdgcn -mcpu=gfx1200 %s -o - | FileCheck %s --check-prefixes=GFX12,GISEL-GFX12
+
+; (workitem_id_x | workitem_id_y | workitem_id_z) == 0
+define i1 @workitem_zero() {
+; DAGISEL-GFX9-LABEL: workitem_zero:
+; DAGISEL-GFX9: ; %bb.0: ; %entry
+; DAGISEL-GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; DAGISEL-GFX9-NEXT: v_lshrrev_b32_e32 v1, 10, v31
+; DAGISEL-GFX9-NEXT: v_lshrrev_b32_e32 v0, 20, v31
+; DAGISEL-GFX9-NEXT: v_or_b32_e32 v1, v31, v1
+; DAGISEL-GFX9-NEXT: v_or_b32_e32 v0, v1, v0
+; DAGISEL-GFX9-NEXT: v_and_b32_e32 v0, 0x3ff, v0
+; DAGISEL-GFX9-NEXT: v_cmp_eq_u32_e32 vcc, 0, v0
+; DAGISEL-GFX9-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; DAGISEL-GFX9-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX942-LABEL: workitem_zero:
+; DAGISEL-GFX942: ; %bb.0: ; %entry
+; DAGISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; DAGISEL-GFX942-NEXT: v_lshrrev_b32_e32 v0, 20, v31
+; DAGISEL-GFX942-NEXT: v_lshrrev_b32_e32 v1, 10, v31
+; DAGISEL-GFX942-NEXT: v_or3_b32 v0, v31, v1, v0
+; DAGISEL-GFX942-NEXT: v_and_b32_e32 v0, 0x3ff, v0
+; DAGISEL-GFX942-NEXT: v_cmp_eq_u32_e32 vcc, 0, v0
+; DAGISEL-GFX942-NEXT: s_nop 1
+; DAGISEL-GFX942-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; DAGISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX12-LABEL: workitem_zero:
+; DAGISEL-GFX12: ; %bb.0: ; %entry
+; DAGISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; DAGISEL-GFX12-NEXT: v_lshrrev_b32_e32 v0, 20, v31
+; DAGISEL-GFX12-NEXT: v_lshrrev_b32_e32 v1, 10, v31
+; DAGISEL-GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; DAGISEL-GFX12-NEXT: v_or3_b32 v0, v31, v1, v0
+; DAGISEL-GFX12-NEXT: v_and_b32_e32 v0, 0x3ff, v0
+; DAGISEL-GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1)
+; DAGISEL-GFX12-NEXT: v_cmp_eq_u32_e32 vcc_lo, 0, v0
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffd
+; DAGISEL-GFX12-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc_lo
+; DAGISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX8-LABEL: workitem_zero:
+; GISEL-GFX8: ; %bb.0: ; %entry
+; GISEL-GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GISEL-GFX8-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GISEL-GFX8-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GISEL-GFX8-NEXT: v_or_b32_e32 v0, v0, v1
+; GISEL-GFX8-NEXT: v_bfe_u32 v1, v31, 20, 10
+; GISEL-GFX8-NEXT: v_or_b32_e32 v0, v0, v1
+; GISEL-GFX8-NEXT: v_cmp_eq_u32_e32 vcc, 0, v0
+; GISEL-GFX8-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; GISEL-GFX8-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX942-LABEL: workitem_zero:
+; GISEL-GFX942: ; %bb.0: ; %entry
+; GISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GISEL-GFX942-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GISEL-GFX942-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GISEL-GFX942-NEXT: v_bfe_u32 v2, v31, 20, 10
+; GISEL-GFX942-NEXT: v_or3_b32 v0, v0, v1, v2
+; GISEL-GFX942-NEXT: v_cmp_eq_u32_e32 vcc, 0, v0
+; GISEL-GFX942-NEXT: s_nop 1
+; GISEL-GFX942-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; GISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX12-LABEL: workitem_zero:
+; GISEL-GFX12: ; %bb.0: ; %entry
+; GISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; GISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; GISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; GISEL-GFX12-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GISEL-GFX12-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GISEL-GFX12-NEXT: v_bfe_u32 v2, v31, 20, 10
+; GISEL-GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; GISEL-GFX12-NEXT: v_or3_b32 v0, v0, v1, v2
+; GISEL-GFX12-NEXT: v_cmp_eq_u32_e32 vcc_lo, 0, v0
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffd
+; GISEL-GFX12-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc_lo
+; GISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+entry:
+ %0 = tail call i32 @llvm.amdgcn.workitem.id.x()
+ %1 = tail call i32 @llvm.amdgcn.workitem.id.y()
+ %or = or i32 %0, %1
+ %2 = tail call i32 @llvm.amdgcn.workitem.id.z()
+ %or1 = or i32 %or, %2
+ %cmp = icmp eq i32 %or1, 0
+ ret i1 %cmp
+}
+
+; (workitem_id_x | workitem_id_y | workitem_id_z) != 0
+define i1 @workitem_nonzero() {
+; DAGISEL-GFX9-LABEL: workitem_nonzero:
+; DAGISEL-GFX9: ; %bb.0: ; %entry
+; DAGISEL-GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; DAGISEL-GFX9-NEXT: v_lshrrev_b32_e32 v1, 10, v31
+; DAGISEL-GFX9-NEXT: v_lshrrev_b32_e32 v0, 20, v31
+; DAGISEL-GFX9-NEXT: v_or_b32_e32 v1, v31, v1
+; DAGISEL-GFX9-NEXT: v_or_b32_e32 v0, v1, v0
+; DAGISEL-GFX9-NEXT: v_and_b32_e32 v0, 0x3ff, v0
+; DAGISEL-GFX9-NEXT: v_cmp_ne_u32_e32 vcc, 0, v0
+; DAGISEL-GFX9-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; DAGISEL-GFX9-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX942-LABEL: workitem_nonzero:
+; DAGISEL-GFX942: ; %bb.0: ; %entry
+; DAGISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; DAGISEL-GFX942-NEXT: v_lshrrev_b32_e32 v0, 20, v31
+; DAGISEL-GFX942-NEXT: v_lshrrev_b32_e32 v1, 10, v31
+; DAGISEL-GFX942-NEXT: v_or3_b32 v0, v31, v1, v0
+; DAGISEL-GFX942-NEXT: v_and_b32_e32 v0, 0x3ff, v0
+; DAGISEL-GFX942-NEXT: v_cmp_ne_u32_e32 vcc, 0, v0
+; DAGISEL-GFX942-NEXT: s_nop 1
+; DAGISEL-GFX942-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; DAGISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX12-LABEL: workitem_nonzero:
+; DAGISEL-GFX12: ; %bb.0: ; %entry
+; DAGISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; DAGISEL-GFX12-NEXT: v_lshrrev_b32_e32 v0, 20, v31
+; DAGISEL-GFX12-NEXT: v_lshrrev_b32_e32 v1, 10, v31
+; DAGISEL-GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; DAGISEL-GFX12-NEXT: v_or3_b32 v0, v31, v1, v0
+; DAGISEL-GFX12-NEXT: v_and_b32_e32 v0, 0x3ff, v0
+; DAGISEL-GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1)
+; DAGISEL-GFX12-NEXT: v_cmp_ne_u32_e32 vcc_lo, 0, v0
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffd
+; DAGISEL-GFX12-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc_lo
+; DAGISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX8-LABEL: workitem_nonzero:
+; GISEL-GFX8: ; %bb.0: ; %entry
+; GISEL-GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GISEL-GFX8-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GISEL-GFX8-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GISEL-GFX8-NEXT: v_or_b32_e32 v0, v0, v1
+; GISEL-GFX8-NEXT: v_bfe_u32 v1, v31, 20, 10
+; GISEL-GFX8-NEXT: v_or_b32_e32 v0, v0, v1
+; GISEL-GFX8-NEXT: v_cmp_ne_u32_e32 vcc, 0, v0
+; GISEL-GFX8-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; GISEL-GFX8-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX942-LABEL: workitem_nonzero:
+; GISEL-GFX942: ; %bb.0: ; %entry
+; GISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GISEL-GFX942-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GISEL-GFX942-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GISEL-GFX942-NEXT: v_bfe_u32 v2, v31, 20, 10
+; GISEL-GFX942-NEXT: v_or3_b32 v0, v0, v1, v2
+; GISEL-GFX942-NEXT: v_cmp_ne_u32_e32 vcc, 0, v0
+; GISEL-GFX942-NEXT: s_nop 1
+; GISEL-GFX942-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; GISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX12-LABEL: workitem_nonzero:
+; GISEL-GFX12: ; %bb.0: ; %entry
+; GISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; GISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; GISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; GISEL-GFX12-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GISEL-GFX12-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GISEL-GFX12-NEXT: v_bfe_u32 v2, v31, 20, 10
+; GISEL-GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; GISEL-GFX12-NEXT: v_or3_b32 v0, v0, v1, v2
+; GISEL-GFX12-NEXT: v_cmp_ne_u32_e32 vcc_lo, 0, v0
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffd
+; GISEL-GFX12-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc_lo
+; GISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+entry:
+ %0 = tail call i32 @llvm.amdgcn.workitem.id.x()
+ %1 = tail call i32 @llvm.amdgcn.workitem.id.y()
+ %or = or i32 %0, %1
+ %2 = tail call i32 @llvm.amdgcn.workitem.id.z()
+ %or1 = or i32 %or, %2
+ %cmp = icmp ne i32 %or1, 0
+ ret i1 %cmp
+}
+
+; (workgroup_id_x | workgroup_id_y | workgroup_id_z) == 0
+define i1 @workgroup_zero() {
+; DAGISEL-GFX9-LABEL: workgroup_zero:
+; DAGISEL-GFX9: ; %bb.0: ; %entry
+; DAGISEL-GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; DAGISEL-GFX9-NEXT: s_or_b32 s4, s12, s13
+; DAGISEL-GFX9-NEXT: s_or_b32 s4, s4, s14
+; DAGISEL-GFX9-NEXT: s_cmp_eq_u32 s4, 0
+; DAGISEL-GFX9-NEXT: s_cselect_b64 s[4:5], -1, 0
+; DAGISEL-GFX9-NEXT: v_cndmask_b32_e64 v0, 0, 1, s[4:5]
+; DAGISEL-GFX9-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX942-LABEL: workgroup_zero:
+; DAGISEL-GFX942: ; %bb.0: ; %entry
+; DAGISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; DAGISEL-GFX942-NEXT: s_or_b32 s0, s12, s13
+; DAGISEL-GFX942-NEXT: s_or_b32 s0, s0, s14
+; DAGISEL-GFX942-NEXT: s_cmp_eq_u32 s0, 0
+; DAGISEL-GFX942-NEXT: s_cselect_b64 s[0:1], -1, 0
+; DAGISEL-GFX942-NEXT: v_cndmask_b32_e64 v0, 0, 1, s[0:1]
+; DAGISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX12-LABEL: workgroup_zero:
+; DAGISEL-GFX12: ; %bb.0: ; %entry
+; DAGISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; DAGISEL-GFX12-NEXT: s_and_b32 s0, ttmp7, 0xffff
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: s_lshr_b32 s1, ttmp7, 16
+; DAGISEL-GFX12-NEXT: s_or_b32 s0, ttmp9, s0
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: s_or_b32 s0, s0, s1
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: s_cmp_eq_u32 s0, 0
+; DAGISEL-GFX12-NEXT: s_cselect_b32 s0, -1, 0
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: v_cndmask_b32_e64 v0, 0, 1, s0
+; DAGISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX8-LABEL: workgroup_zero:
+; GISEL-GFX8: ; %bb.0: ; %entry
+; GISEL-GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GISEL-GFX8-NEXT: s_or_b32 s4, s12, s13
+; GISEL-GFX8-NEXT: s_or_b32 s4, s4, s14
+; GISEL-GFX8-NEXT: s_cmp_eq_u32 s4, 0
+; GISEL-GFX8-NEXT: s_cselect_b32 s4, 1, 0
+; GISEL-GFX8-NEXT: v_mov_b32_e32 v0, s4
+; GISEL-GFX8-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX942-LABEL: workgroup_zero:
+; GISEL-GFX942: ; %bb.0: ; %entry
+; GISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GISEL-GFX942-NEXT: s_or_b32 s0, s12, s13
+; GISEL-GFX942-NEXT: s_or_b32 s0, s0, s14
+; GISEL-GFX942-NEXT: s_cmp_eq_u32 s0, 0
+; GISEL-GFX942-NEXT: s_cselect_b32 s0, 1, 0
+; GISEL-GFX942-NEXT: v_mov_b32_e32 v0, s0
+; GISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX12-LABEL: workgroup_zero:
+; GISEL-GFX12: ; %bb.0: ; %entry
+; GISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; GISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; GISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; GISEL-GFX12-NEXT: s_and_b32 s0, ttmp7, 0xffff
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: s_lshr_b32 s1, ttmp7, 16
+; GISEL-GFX12-NEXT: s_or_b32 s0, ttmp9, s0
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: s_or_b32 s0, s0, s1
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: s_cmp_eq_u32 s0, 0
+; GISEL-GFX12-NEXT: s_cselect_b32 s0, 1, 0
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: v_mov_b32_e32 v0, s0
+; GISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+entry:
+ %0 = tail call i32 @llvm.amdgcn.workgroup.id.x()
+ %1 = tail call i32 @llvm.amdgcn.workgroup.id.y()
+ %or = or i32 %0, %1
+ %2 = tail call i32 @llvm.amdgcn.workgroup.id.z()
+ %or1 = or i32 %or, %2
+ %cmp = icmp eq i32 %or1, 0
+ ret i1 %cmp
+}
+
+; (workgroup_id_x | workgroup_id_y | workgroup_id_z) != 0
+define i1 @workgroup_nonzero() {
+; DAGISEL-GFX9-LABEL: workgroup_nonzero:
+; DAGISEL-GFX9: ; %bb.0: ; %entry
+; DAGISEL-GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; DAGISEL-GFX9-NEXT: s_or_b32 s4, s12, s13
+; DAGISEL-GFX9-NEXT: s_or_b32 s4, s4, s14
+; DAGISEL-GFX9-NEXT: s_cmp_lg_u32 s4, 0
+; DAGISEL-GFX9-NEXT: s_cselect_b64 s[4:5], -1, 0
+; DAGISEL-GFX9-NEXT: v_cndmask_b32_e64 v0, 0, 1, s[4:5]
+; DAGISEL-GFX9-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX942-LABEL: workgroup_nonzero:
+; DAGISEL-GFX942: ; %bb.0: ; %entry
+; DAGISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; DAGISEL-GFX942-NEXT: s_or_b32 s0, s12, s13
+; DAGISEL-GFX942-NEXT: s_or_b32 s0, s0, s14
+; DAGISEL-GFX942-NEXT: s_cmp_lg_u32 s0, 0
+; DAGISEL-GFX942-NEXT: s_cselect_b64 s[0:1], -1, 0
+; DAGISEL-GFX942-NEXT: v_cndmask_b32_e64 v0, 0, 1, s[0:1]
+; DAGISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX12-LABEL: workgroup_nonzero:
+; DAGISEL-GFX12: ; %bb.0: ; %entry
+; DAGISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; DAGISEL-GFX12-NEXT: s_and_b32 s0, ttmp7, 0xffff
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: s_lshr_b32 s1, ttmp7, 16
+; DAGISEL-GFX12-NEXT: s_or_b32 s0, ttmp9, s0
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: s_or_b32 s0, s0, s1
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: s_cmp_lg_u32 s0, 0
+; DAGISEL-GFX12-NEXT: s_cselect_b32 s0, -1, 0
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: v_cndmask_b32_e64 v0, 0, 1, s0
+; DAGISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX8-LABEL: workgroup_nonzero:
+; GISEL-GFX8: ; %bb.0: ; %entry
+; GISEL-GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GISEL-GFX8-NEXT: s_or_b32 s4, s12, s13
+; GISEL-GFX8-NEXT: s_or_b32 s4, s4, s14
+; GISEL-GFX8-NEXT: s_cmp_lg_u32 s4, 0
+; GISEL-GFX8-NEXT: s_cselect_b32 s4, 1, 0
+; GISEL-GFX8-NEXT: v_mov_b32_e32 v0, s4
+; GISEL-GFX8-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX942-LABEL: workgroup_nonzero:
+; GISEL-GFX942: ; %bb.0: ; %entry
+; GISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GISEL-GFX942-NEXT: s_or_b32 s0, s12, s13
+; GISEL-GFX942-NEXT: s_or_b32 s0, s0, s14
+; GISEL-GFX942-NEXT: s_cmp_lg_u32 s0, 0
+; GISEL-GFX942-NEXT: s_cselect_b32 s0, 1, 0
+; GISEL-GFX942-NEXT: v_mov_b32_e32 v0, s0
+; GISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX12-LABEL: workgroup_nonzero:
+; GISEL-GFX12: ; %bb.0: ; %entry
+; GISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; GISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; GISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; GISEL-GFX12-NEXT: s_and_b32 s0, ttmp7, 0xffff
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: s_lshr_b32 s1, ttmp7, 16
+; GISEL-GFX12-NEXT: s_or_b32 s0, ttmp9, s0
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: s_or_b32 s0, s0, s1
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: s_cmp_lg_u32 s0, 0
+; GISEL-GFX12-NEXT: s_cselect_b32 s0, 1, 0
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: v_mov_b32_e32 v0, s0
+; GISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+entry:
+ %0 = tail call i32 @llvm.amdgcn.workgroup.id.x()
+ %1 = tail call i32 @llvm.amdgcn.workgroup.id.y()
+ %or = or i32 %0, %1
+ %2 = tail call i32 @llvm.amdgcn.workgroup.id.z()
+ %or1 = or i32 %or, %2
+ %cmp = icmp ne i32 %or1, 0
+ ret i1 %cmp
+}
+
+; (workitem_id_x | workitem_id_y | workitem_id_z | workgroup_id_x | workgroup_id_y | workgroup_id_z) == 0
+define i1 @workitem_workgroup_zero() {
+; GFX8-LABEL: workitem_workgroup_zero:
+; GFX8: ; %bb.0: ; %entry
+; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8-NEXT: s_or_b32 s4, s12, s13
+; GFX8-NEXT: s_or_b32 s4, s4, s14
+; GFX8-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GFX8-NEXT: v_or_b32_e32 v0, s4, v0
+; GFX8-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GFX8-NEXT: v_or_b32_e32 v0, v0, v1
+; GFX8-NEXT: v_cmp_eq_u32_e32 vcc, 0, v0
+; GFX8-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; GFX8-NEXT: s_setpc_b64 s[30:31]
+;
+; GFX942-LABEL: workitem_workgroup_zero:
+; GFX942: ; %bb.0: ; %entry
+; GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX942-NEXT: s_or_b32 s0, s12, s13
+; GFX942-NEXT: s_or_b32 s0, s0, s14
+; GFX942-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GFX942-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GFX942-NEXT: v_or3_b32 v0, s0, v0, v1
+; GFX942-NEXT: v_cmp_eq_u32_e32 vcc, 0, v0
+; GFX942-NEXT: s_nop 1
+; GFX942-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; GFX12-LABEL: workitem_workgroup_zero:
+; GFX12: ; %bb.0: ; %entry
+; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; GFX12-NEXT: s_wait_expcnt 0x0
+; GFX12-NEXT: s_wait_samplecnt 0x0
+; GFX12-NEXT: s_wait_bvhcnt 0x0
+; GFX12-NEXT: s_wait_kmcnt 0x0
+; GFX12-NEXT: s_and_b32 s0, ttmp7, 0xffff
+; GFX12-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GFX12-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GFX12-NEXT: s_wait_alu 0xfffe
+; GFX12-NEXT: s_lshr_b32 s1, ttmp7, 16
+; GFX12-NEXT: s_or_b32 s0, ttmp9, s0
+; GFX12-NEXT: s_wait_alu 0xfffe
+; GFX12-NEXT: s_or_b32 s0, s0, s1
+; GFX12-NEXT: s_wait_alu 0xfffe
+; GFX12-NEXT: v_or3_b32 v0, s0, v0, v1
+; GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1)
+; GFX12-NEXT: v_cmp_eq_u32_e32 vcc_lo, 0, v0
+; GFX12-NEXT: s_wait_alu 0xfffd
+; GFX12-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc_lo
+; GFX12-NEXT: s_setpc_b64 s[30:31]
+entry:
+ %0 = tail call i32 @llvm.amdgcn.workgroup.id.x()
+ %1 = tail call i32 @llvm.amdgcn.workgroup.id.y()
+ %or = or i32 %0, %1
+ %2 = tail call i32 @llvm.amdgcn.workgroup.id.z()
+ %or1 = or i32 %or, %2
+ %3 = tail call i32 @llvm.amdgcn.workitem.id.x()
+ %or2 = or i32 %or1, %3
+ %4 = tail call i32 @llvm.amdgcn.workitem.id.y()
+ %or3 = or i32 %or2, %4
+ %5 = tail call i32 @llvm.amdgcn.workitem.id.z()
+ %or4 = or i32 %or3, %5
+ %cmp = icmp eq i32 %or3, 0
+ ret i1 %cmp
+}
+
+; (workitem_id_x | workitem_id_y | workitem_id_z | workgroup_id_x | workgroup_id_y | workgroup_id_z) != 0
+define i1 @workitem_workgroup_nonzero() {
+; GFX8-LABEL: workitem_workgroup_nonzero:
+; GFX8: ; %bb.0: ; %entry
+; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8-NEXT: s_or_b32 s4, s12, s13
+; GFX8-NEXT: s_or_b32 s4, s4, s14
+; GFX8-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; GFX8-NEXT: v_or_b32_e32 v0, s4, v0
+; GFX8-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GFX8-NEXT: v_or_b32_e32 v0, v0, v1
+; GFX8-NEXT: v_bfe_u32 v1, v31, 20, 10
+; GFX8-NEXT: v_or_b32_e32 v0, v0, v1
+; GFX8-NEXT: v_cmp_ne_u32_e32 vcc, 0, v0
+; GFX8-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; GFX8-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX942-LABEL: workitem_workgroup_nonzero:
+; DAGISEL-GFX942: ; %bb.0: ; %entry
+; DAGISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; DAGISEL-GFX942-NEXT: s_or_b32 s0, s12, s13
+; DAGISEL-GFX942-NEXT: s_or_b32 s0, s0, s14
+; DAGISEL-GFX942-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; DAGISEL-GFX942-NEXT: v_or_b32_e32 v0, s0, v0
+; DAGISEL-GFX942-NEXT: v_bfe_u32 v1, v31, 20, 10
+; DAGISEL-GFX942-NEXT: v_bfe_u32 v2, v31, 10, 10
+; DAGISEL-GFX942-NEXT: v_or3_b32 v0, v0, v2, v1
+; DAGISEL-GFX942-NEXT: v_cmp_ne_u32_e32 vcc, 0, v0
+; DAGISEL-GFX942-NEXT: s_nop 1
+; DAGISEL-GFX942-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; DAGISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; DAGISEL-GFX12-LABEL: workitem_workgroup_nonzero:
+; DAGISEL-GFX12: ; %bb.0: ; %entry
+; DAGISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; DAGISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; DAGISEL-GFX12-NEXT: v_and_b32_e32 v0, 0x3ff, v31
+; DAGISEL-GFX12-NEXT: s_and_b32 s0, ttmp7, 0xffff
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: s_lshr_b32 s1, ttmp7, 16
+; DAGISEL-GFX12-NEXT: s_or_b32 s0, ttmp9, s0
+; DAGISEL-GFX12-NEXT: v_bfe_u32 v1, v31, 20, 10
+; DAGISEL-GFX12-NEXT: v_bfe_u32 v2, v31, 10, 10
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; DAGISEL-GFX12-NEXT: v_or3_b32 v0, s0, s1, v0
+; DAGISEL-GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; DAGISEL-GFX12-NEXT: v_or3_b32 v0, v0, v2, v1
+; DAGISEL-GFX12-NEXT: v_cmp_ne_u32_e32 vcc_lo, 0, v0
+; DAGISEL-GFX12-NEXT: s_wait_alu 0xfffd
+; DAGISEL-GFX12-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc_lo
+; DAGISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX942-LABEL: workitem_workgroup_nonzero:
+; GISEL-GFX942: ; %bb.0: ; %entry
+; GISEL-GFX942-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GISEL-GFX942-NEXT: s_or_b32 s0, s12, s13
+; GISEL-GFX942-NEXT: s_or_b32 s0, s0, s14
+; GISEL-GFX942-NEXT: v_mov_b32_e32 v0, 0x3ff
+; GISEL-GFX942-NEXT: v_and_or_b32 v0, v31, v0, s0
+; GISEL-GFX942-NEXT: v_bfe_u32 v1, v31, 10, 10
+; GISEL-GFX942-NEXT: v_bfe_u32 v2, v31, 20, 10
+; GISEL-GFX942-NEXT: v_or3_b32 v0, v0, v1, v2
+; GISEL-GFX942-NEXT: v_cmp_ne_u32_e32 vcc, 0, v0
+; GISEL-GFX942-NEXT: s_nop 1
+; GISEL-GFX942-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc
+; GISEL-GFX942-NEXT: s_setpc_b64 s[30:31]
+;
+; GISEL-GFX12-LABEL: workitem_workgroup_nonzero:
+; GISEL-GFX12: ; %bb.0: ; %entry
+; GISEL-GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
+; GISEL-GFX12-NEXT: s_wait_expcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_samplecnt 0x0
+; GISEL-GFX12-NEXT: s_wait_bvhcnt 0x0
+; GISEL-GFX12-NEXT: s_wait_kmcnt 0x0
+; GISEL-GFX12-NEXT: s_and_b32 s0, ttmp7, 0xffff
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: s_lshr_b32 s1, ttmp7, 16
+; GISEL-GFX12-NEXT: s_or_b32 s0, ttmp9, s0
+; GISEL-GFX12-NEXT: v_bfe_u32 v0, v31, 10, 10
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: s_or_b32 s0, s0, s1
+; GISEL-GFX12-NEXT: v_bfe_u32 v1, v31, 20, 10
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffe
+; GISEL-GFX12-NEXT: v_and_or_b32 v2, 0x3ff, v31, s0
+; GISEL-GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1) | instskip(NEXT) | instid1(VALU_DEP_1)
+; GISEL-GFX12-NEXT: v_or3_b32 v0, v2, v0, v1
+; GISEL-GFX12-NEXT: v_cmp_ne_u32_e32 vcc_lo, 0, v0
+; GISEL-GFX12-NEXT: s_wait_alu 0xfffd
+; GISEL-GFX12-NEXT: v_cndmask_b32_e64 v0, 0, 1, vcc_lo
+; GISEL-GFX12-NEXT: s_setpc_b64 s[30:31]
+entry:
+ %0 = tail call i32 @llvm.amdgcn.workgroup.id.x()
+ %1 = tail call i32 @llvm.amdgcn.workgroup.id.y()
+ %or = or i32 %0, %1
+ %2 = tail call i32 @llvm.amdgcn.workgroup.id.z()
+ %or1 = or i32 %or, %2
+ %3 = tail call i32 @llvm.amdgcn.workitem.id.x()
+ %or2 = or i32 %or1, %3
+ %4 = tail call i32 @llvm.amdgcn.workitem.id.y()
+ %or3 = or i32 %or2, %4
+ %5 = tail call i32 @llvm.amdgcn.workitem.id.z()
+ %or4 = or i32 %or3, %5
+ %cmp = icmp ne i32 %or4, 0
+ ret i1 %cmp
+}
More information about the llvm-branch-commits
mailing list