[llvm] [NFC][AMDGPU] Update and.ll test and automate check line generation (PR #145371)
via llvm-commits
llvm-commits at lists.llvm.org
Mon Jun 23 10:34:29 PDT 2025
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-amdgpu
Author: Chris Jackson (chrisjbris)
<details>
<summary>Changes</summary>
- Convert the test to use update_llc_test_checks.py.
- Use different check prefixes for the different -mcpu options.
- Remove unused xnack 'off' flag.
---
Patch is 96.46 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/145371.diff
1 Files Affected:
- (modified) llvm/test/CodeGen/AMDGPU/and.ll (+1974-247)
``````````diff
diff --git a/llvm/test/CodeGen/AMDGPU/and.ll b/llvm/test/CodeGen/AMDGPU/and.ll
index c6233642110ea..e5fe9195e2dcc 100644
--- a/llvm/test/CodeGen/AMDGPU/and.ll
+++ b/llvm/test/CodeGen/AMDGPU/and.ll
@@ -1,17 +1,58 @@
-; RUN: llc -mtriple=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -mtriple=amdgcn -mcpu=tonga -mattr=-flat-for-global,-xnack -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -mtriple=r600 -mcpu=redwood < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -mtriple=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GFX6 %s
+; RUN: llc -mtriple=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GFX8 %s
+; RUN: llc -mtriple=r600 -mcpu=redwood < %s | FileCheck -check-prefix=EG %s
declare i32 @llvm.amdgcn.workitem.id.x() #0
-; FUNC-LABEL: {{^}}test2:
-; EG: AND_INT {{\*? *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; EG: AND_INT {{\*? *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-
-; SI: s_and_b32 s{{[0-9]+, s[0-9]+, s[0-9]+}}
-; SI: s_and_b32 s{{[0-9]+, s[0-9]+, s[0-9]+}}
-
define amdgpu_kernel void @test2(ptr addrspace(1) %out, ptr addrspace(1) %in) {
+; GFX6-LABEL: test2:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x9
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: s_load_dwordx4 s[4:7], s[2:3], 0x0
+; GFX6-NEXT: s_mov_b32 s3, 0xf000
+; GFX6-NEXT: s_mov_b32 s2, -1
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: s_and_b32 s5, s5, s7
+; GFX6-NEXT: s_and_b32 s4, s4, s6
+; GFX6-NEXT: v_mov_b32_e32 v0, s4
+; GFX6-NEXT: v_mov_b32_e32 v1, s5
+; GFX6-NEXT: buffer_store_dwordx2 v[0:1], off, s[0:3], 0
+; GFX6-NEXT: s_endpgm
+;
+; GFX8-LABEL: test2:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: s_load_dwordx4 s[4:7], s[2:3], 0x0
+; GFX8-NEXT: s_mov_b32 s3, 0xf000
+; GFX8-NEXT: s_mov_b32 s2, -1
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: s_and_b32 s5, s5, s7
+; GFX8-NEXT: s_and_b32 s4, s4, s6
+; GFX8-NEXT: v_mov_b32_e32 v0, s4
+; GFX8-NEXT: v_mov_b32_e32 v1, s5
+; GFX8-NEXT: buffer_store_dwordx2 v[0:1], off, s[0:3], 0
+; GFX8-NEXT: s_endpgm
+;
+; EG-LABEL: test2:
+; EG: ; %bb.0:
+; EG-NEXT: ALU 0, @8, KC0[CB0:0-32], KC1[]
+; EG-NEXT: TEX 0 @6
+; EG-NEXT: ALU 3, @9, KC0[CB0:0-32], KC1[]
+; EG-NEXT: MEM_RAT_CACHELESS STORE_RAW T0.XY, T1.X, 1
+; EG-NEXT: CF_END
+; EG-NEXT: PAD
+; EG-NEXT: Fetch clause starting at 6:
+; EG-NEXT: VTX_READ_128 T0.XYZW, T0.X, 0, #1
+; EG-NEXT: ALU clause starting at 8:
+; EG-NEXT: MOV * T0.X, KC0[2].Z,
+; EG-NEXT: ALU clause starting at 9:
+; EG-NEXT: AND_INT * T0.Y, T0.Y, T0.W,
+; EG-NEXT: AND_INT T0.X, T0.X, T0.Z,
+; EG-NEXT: LSHR * T1.X, KC0[2].Y, literal.x,
+; EG-NEXT: 2(2.802597e-45), 0(0.000000e+00)
%b_ptr = getelementptr <2 x i32>, ptr addrspace(1) %in, i32 1
%a = load <2 x i32>, ptr addrspace(1) %in
%b = load <2 x i32>, ptr addrspace(1) %b_ptr
@@ -20,19 +61,65 @@ define amdgpu_kernel void @test2(ptr addrspace(1) %out, ptr addrspace(1) %in) {
ret void
}
-; FUNC-LABEL: {{^}}test4:
-; EG: AND_INT {{\** *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; EG: AND_INT {{\** *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; EG: AND_INT {{\** *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; EG: AND_INT {{\** *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-
-
-; SI: s_and_b32 s{{[0-9]+, s[0-9]+, s[0-9]+}}
-; SI: s_and_b32 s{{[0-9]+, s[0-9]+, s[0-9]+}}
-; SI: s_and_b32 s{{[0-9]+, s[0-9]+, s[0-9]+}}
-; SI: s_and_b32 s{{[0-9]+, s[0-9]+, s[0-9]+}}
-
define amdgpu_kernel void @test4(ptr addrspace(1) %out, ptr addrspace(1) %in) {
+; GFX6-LABEL: test4:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_load_dwordx4 s[8:11], s[4:5], 0x9
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: s_load_dwordx8 s[0:7], s[10:11], 0x0
+; GFX6-NEXT: s_mov_b32 s11, 0xf000
+; GFX6-NEXT: s_mov_b32 s10, -1
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: s_and_b32 s3, s3, s7
+; GFX6-NEXT: s_and_b32 s2, s2, s6
+; GFX6-NEXT: s_and_b32 s1, s1, s5
+; GFX6-NEXT: s_and_b32 s0, s0, s4
+; GFX6-NEXT: v_mov_b32_e32 v0, s0
+; GFX6-NEXT: v_mov_b32_e32 v1, s1
+; GFX6-NEXT: v_mov_b32_e32 v2, s2
+; GFX6-NEXT: v_mov_b32_e32 v3, s3
+; GFX6-NEXT: buffer_store_dwordx4 v[0:3], off, s[8:11], 0
+; GFX6-NEXT: s_endpgm
+;
+; GFX8-LABEL: test4:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_load_dwordx4 s[8:11], s[4:5], 0x24
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: s_load_dwordx8 s[0:7], s[10:11], 0x0
+; GFX8-NEXT: s_mov_b32 s11, 0xf000
+; GFX8-NEXT: s_mov_b32 s10, -1
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: s_and_b32 s3, s3, s7
+; GFX8-NEXT: s_and_b32 s2, s2, s6
+; GFX8-NEXT: s_and_b32 s1, s1, s5
+; GFX8-NEXT: s_and_b32 s0, s0, s4
+; GFX8-NEXT: v_mov_b32_e32 v0, s0
+; GFX8-NEXT: v_mov_b32_e32 v1, s1
+; GFX8-NEXT: v_mov_b32_e32 v2, s2
+; GFX8-NEXT: v_mov_b32_e32 v3, s3
+; GFX8-NEXT: buffer_store_dwordx4 v[0:3], off, s[8:11], 0
+; GFX8-NEXT: s_endpgm
+;
+; EG-LABEL: test4:
+; EG: ; %bb.0:
+; EG-NEXT: ALU 0, @10, KC0[CB0:0-32], KC1[]
+; EG-NEXT: TEX 1 @6
+; EG-NEXT: ALU 5, @11, KC0[CB0:0-32], KC1[]
+; EG-NEXT: MEM_RAT_CACHELESS STORE_RAW T0.XYZW, T1.X, 1
+; EG-NEXT: CF_END
+; EG-NEXT: PAD
+; EG-NEXT: Fetch clause starting at 6:
+; EG-NEXT: VTX_READ_128 T1.XYZW, T0.X, 16, #1
+; EG-NEXT: VTX_READ_128 T0.XYZW, T0.X, 0, #1
+; EG-NEXT: ALU clause starting at 10:
+; EG-NEXT: MOV * T0.X, KC0[2].Z,
+; EG-NEXT: ALU clause starting at 11:
+; EG-NEXT: AND_INT * T0.W, T0.W, T1.W,
+; EG-NEXT: AND_INT * T0.Z, T0.Z, T1.Z,
+; EG-NEXT: AND_INT * T0.Y, T0.Y, T1.Y,
+; EG-NEXT: AND_INT T0.X, T0.X, T1.X,
+; EG-NEXT: LSHR * T1.X, KC0[2].Y, literal.x,
+; EG-NEXT: 2(2.802597e-45), 0(0.000000e+00)
%b_ptr = getelementptr <4 x i32>, ptr addrspace(1) %in, i32 1
%a = load <4 x i32>, ptr addrspace(1) %in
%b = load <4 x i32>, ptr addrspace(1) %b_ptr
@@ -41,17 +128,83 @@ define amdgpu_kernel void @test4(ptr addrspace(1) %out, ptr addrspace(1) %in) {
ret void
}
-; FUNC-LABEL: {{^}}s_and_i32:
-; SI: s_and_b32
define amdgpu_kernel void @s_and_i32(ptr addrspace(1) %out, i32 %a, i32 %b) {
+; GFX6-LABEL: s_and_i32:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x9
+; GFX6-NEXT: s_mov_b32 s7, 0xf000
+; GFX6-NEXT: s_mov_b32 s6, -1
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: s_mov_b32 s4, s0
+; GFX6-NEXT: s_and_b32 s0, s2, s3
+; GFX6-NEXT: s_mov_b32 s5, s1
+; GFX6-NEXT: v_mov_b32_e32 v0, s0
+; GFX6-NEXT: buffer_store_dword v0, off, s[4:7], 0
+; GFX6-NEXT: s_endpgm
+;
+; GFX8-LABEL: s_and_i32:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8-NEXT: s_mov_b32 s7, 0xf000
+; GFX8-NEXT: s_mov_b32 s6, -1
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: s_mov_b32 s4, s0
+; GFX8-NEXT: s_and_b32 s0, s2, s3
+; GFX8-NEXT: s_mov_b32 s5, s1
+; GFX8-NEXT: v_mov_b32_e32 v0, s0
+; GFX8-NEXT: buffer_store_dword v0, off, s[4:7], 0
+; GFX8-NEXT: s_endpgm
+;
+; EG-LABEL: s_and_i32:
+; EG: ; %bb.0:
+; EG-NEXT: ALU 2, @4, KC0[CB0:0-32], KC1[]
+; EG-NEXT: MEM_RAT_CACHELESS STORE_RAW T1.X, T0.X, 1
+; EG-NEXT: CF_END
+; EG-NEXT: PAD
+; EG-NEXT: ALU clause starting at 4:
+; EG-NEXT: LSHR T0.X, KC0[2].Y, literal.x,
+; EG-NEXT: AND_INT * T1.X, KC0[2].Z, KC0[2].W,
+; EG-NEXT: 2(2.802597e-45), 0(0.000000e+00)
%and = and i32 %a, %b
store i32 %and, ptr addrspace(1) %out, align 4
ret void
}
-; FUNC-LABEL: {{^}}s_and_constant_i32:
-; SI: s_and_b32 s{{[0-9]+}}, s{{[0-9]+}}, 0x12d687
define amdgpu_kernel void @s_and_constant_i32(ptr addrspace(1) %out, i32 %a) {
+; GFX6-LABEL: s_and_constant_i32:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_load_dword s6, s[4:5], 0xb
+; GFX6-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x9
+; GFX6-NEXT: s_mov_b32 s3, 0xf000
+; GFX6-NEXT: s_mov_b32 s2, -1
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: s_and_b32 s4, s6, 0x12d687
+; GFX6-NEXT: v_mov_b32_e32 v0, s4
+; GFX6-NEXT: buffer_store_dword v0, off, s[0:3], 0
+; GFX6-NEXT: s_endpgm
+;
+; GFX8-LABEL: s_and_constant_i32:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_load_dword s6, s[4:5], 0x2c
+; GFX8-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8-NEXT: s_mov_b32 s3, 0xf000
+; GFX8-NEXT: s_mov_b32 s2, -1
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: s_and_b32 s4, s6, 0x12d687
+; GFX8-NEXT: v_mov_b32_e32 v0, s4
+; GFX8-NEXT: buffer_store_dword v0, off, s[0:3], 0
+; GFX8-NEXT: s_endpgm
+;
+; EG-LABEL: s_and_constant_i32:
+; EG: ; %bb.0:
+; EG-NEXT: ALU 2, @4, KC0[CB0:0-32], KC1[]
+; EG-NEXT: MEM_RAT_CACHELESS STORE_RAW T1.X, T0.X, 1
+; EG-NEXT: CF_END
+; EG-NEXT: PAD
+; EG-NEXT: ALU clause starting at 4:
+; EG-NEXT: LSHR T0.X, KC0[2].Y, literal.x,
+; EG-NEXT: AND_INT * T1.X, KC0[2].Z, literal.y,
+; EG-NEXT: 2(2.802597e-45), 1234567(1.729997e-39)
%and = and i32 %a, 1234567
store i32 %and, ptr addrspace(1) %out, align 4
ret void
@@ -62,11 +215,57 @@ define amdgpu_kernel void @s_and_constant_i32(ptr addrspace(1) %out, i32 %a) {
; directly without copying from the SGPR.
; Second use is a VGPR use of the constant.
-; FUNC-LABEL: {{^}}s_and_multi_use_constant_i32_0:
-; SI-DAG: s_and_b32 [[AND:s[0-9]+]], s{{[0-9]+}}, 0x12d687
-; SI-DAG: v_mov_b32_e32 [[VK:v[0-9]+]], 0x12d687
-; SI: buffer_store_dword [[VK]]
+
define amdgpu_kernel void @s_and_multi_use_constant_i32_0(ptr addrspace(1) %out, i32 %a, i32 %b) {
+; GFX6-LABEL: s_and_multi_use_constant_i32_0:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x9
+; GFX6-NEXT: s_mov_b32 s7, 0xf000
+; GFX6-NEXT: s_mov_b32 s6, -1
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: s_mov_b32 s4, s0
+; GFX6-NEXT: s_and_b32 s0, s2, 0x12d687
+; GFX6-NEXT: s_add_i32 s0, s0, s3
+; GFX6-NEXT: s_mov_b32 s5, s1
+; GFX6-NEXT: v_mov_b32_e32 v0, s0
+; GFX6-NEXT: buffer_store_dword v0, off, s[4:7], 0
+; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0)
+; GFX6-NEXT: v_mov_b32_e32 v0, 0x12d687
+; GFX6-NEXT: buffer_store_dword v0, off, s[4:7], 0
+; GFX6-NEXT: s_waitcnt vmcnt(0)
+; GFX6-NEXT: s_endpgm
+;
+; GFX8-LABEL: s_and_multi_use_constant_i32_0:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8-NEXT: s_mov_b32 s7, 0xf000
+; GFX8-NEXT: s_mov_b32 s6, -1
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: s_mov_b32 s4, s0
+; GFX8-NEXT: s_and_b32 s0, s2, 0x12d687
+; GFX8-NEXT: s_add_i32 s0, s0, s3
+; GFX8-NEXT: s_mov_b32 s5, s1
+; GFX8-NEXT: v_mov_b32_e32 v0, s0
+; GFX8-NEXT: buffer_store_dword v0, off, s[4:7], 0
+; GFX8-NEXT: s_waitcnt vmcnt(0)
+; GFX8-NEXT: v_mov_b32_e32 v0, 0x12d687
+; GFX8-NEXT: buffer_store_dword v0, off, s[4:7], 0
+; GFX8-NEXT: s_waitcnt vmcnt(0)
+; GFX8-NEXT: s_endpgm
+;
+; EG-LABEL: s_and_multi_use_constant_i32_0:
+; EG: ; %bb.0:
+; EG-NEXT: ALU 5, @4, KC0[CB0:0-32], KC1[]
+; EG-NEXT: MEM_RAT_CACHELESS STORE_RAW T1.X, T2.X, 0
+; EG-NEXT: MEM_RAT_CACHELESS STORE_RAW T0.X, T2.X, 1
+; EG-NEXT: CF_END
+; EG-NEXT: ALU clause starting at 4:
+; EG-NEXT: MOV T0.X, literal.x,
+; EG-NEXT: AND_INT * T0.W, KC0[2].Z, literal.x,
+; EG-NEXT: 1234567(1.729997e-39), 0(0.000000e+00)
+; EG-NEXT: ADD_INT T1.X, PV.W, KC0[2].W,
+; EG-NEXT: LSHR * T2.X, KC0[2].Y, literal.x,
+; EG-NEXT: 2(2.802597e-45), 0(0.000000e+00)
%and = and i32 %a, 1234567
; Just to stop future replacement of copy to vgpr + store with VALU op.
@@ -77,13 +276,53 @@ define amdgpu_kernel void @s_and_multi_use_constant_i32_0(ptr addrspace(1) %out,
}
; Second use is another SGPR use of the constant.
-; FUNC-LABEL: {{^}}s_and_multi_use_constant_i32_1:
-; SI: s_and_b32 [[AND:s[0-9]+]], s{{[0-9]+}}, 0x12d687
-; SI: s_add_i32
-; SI: s_add_i32 [[ADD:s[0-9]+]], s{{[0-9]+}}, 0x12d687
-; SI: v_mov_b32_e32 [[VADD:v[0-9]+]], [[ADD]]
-; SI: buffer_store_dword [[VADD]]
+
define amdgpu_kernel void @s_and_multi_use_constant_i32_1(ptr addrspace(1) %out, i32 %a, i32 %b) {
+; GFX6-LABEL: s_and_multi_use_constant_i32_1:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x9
+; GFX6-NEXT: s_mov_b32 s7, 0xf000
+; GFX6-NEXT: s_mov_b32 s6, -1
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: s_mov_b32 s4, s0
+; GFX6-NEXT: s_and_b32 s0, s2, 0x12d687
+; GFX6-NEXT: s_add_i32 s0, s0, s3
+; GFX6-NEXT: s_add_i32 s0, s0, 0x12d687
+; GFX6-NEXT: s_mov_b32 s5, s1
+; GFX6-NEXT: v_mov_b32_e32 v0, s0
+; GFX6-NEXT: buffer_store_dword v0, off, s[4:7], 0
+; GFX6-NEXT: s_waitcnt vmcnt(0)
+; GFX6-NEXT: s_endpgm
+;
+; GFX8-LABEL: s_and_multi_use_constant_i32_1:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8-NEXT: s_mov_b32 s7, 0xf000
+; GFX8-NEXT: s_mov_b32 s6, -1
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: s_mov_b32 s4, s0
+; GFX8-NEXT: s_and_b32 s0, s2, 0x12d687
+; GFX8-NEXT: s_add_i32 s0, s0, s3
+; GFX8-NEXT: s_add_i32 s0, s0, 0x12d687
+; GFX8-NEXT: s_mov_b32 s5, s1
+; GFX8-NEXT: v_mov_b32_e32 v0, s0
+; GFX8-NEXT: buffer_store_dword v0, off, s[4:7], 0
+; GFX8-NEXT: s_waitcnt vmcnt(0)
+; GFX8-NEXT: s_endpgm
+;
+; EG-LABEL: s_and_multi_use_constant_i32_1:
+; EG: ; %bb.0:
+; EG-NEXT: ALU 5, @4, KC0[CB0:0-32], KC1[]
+; EG-NEXT: MEM_RAT_CACHELESS STORE_RAW T0.X, T1.X, 1
+; EG-NEXT: CF_END
+; EG-NEXT: PAD
+; EG-NEXT: ALU clause starting at 4:
+; EG-NEXT: AND_INT * T0.W, KC0[2].Z, literal.x,
+; EG-NEXT: 1234567(1.729997e-39), 0(0.000000e+00)
+; EG-NEXT: ADD_INT * T0.W, PV.W, KC0[2].W,
+; EG-NEXT: ADD_INT T0.X, PV.W, literal.x,
+; EG-NEXT: LSHR * T1.X, KC0[2].Y, literal.y,
+; EG-NEXT: 1234567(1.729997e-39), 2(2.802597e-45)
%and = and i32 %a, 1234567
%foo = add i32 %and, 1234567
%bar = add i32 %foo, %b
@@ -91,9 +330,69 @@ define amdgpu_kernel void @s_and_multi_use_constant_i32_1(ptr addrspace(1) %out,
ret void
}
-; FUNC-LABEL: {{^}}v_and_i32_vgpr_vgpr:
-; SI: v_and_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}
define amdgpu_kernel void @v_and_i32_vgpr_vgpr(ptr addrspace(1) %out, ptr addrspace(1) %aptr, ptr addrspace(1) %bptr) {
+; GFX6-LABEL: v_and_i32_vgpr_vgpr:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x9
+; GFX6-NEXT: s_load_dwordx2 s[8:9], s[4:5], 0xd
+; GFX6-NEXT: s_mov_b32 s7, 0xf000
+; GFX6-NEXT: s_mov_b32 s6, 0
+; GFX6-NEXT: v_lshlrev_b32_e32 v0, 2, v0
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: s_mov_b64 s[4:5], s[2:3]
+; GFX6-NEXT: v_mov_b32_e32 v1, 0
+; GFX6-NEXT: s_mov_b64 s[10:11], s[6:7]
+; GFX6-NEXT: buffer_load_dword v2, v[0:1], s[4:7], 0 addr64
+; GFX6-NEXT: buffer_load_dword v3, v[0:1], s[8:11], 0 addr64
+; GFX6-NEXT: s_mov_b64 s[2:3], s[6:7]
+; GFX6-NEXT: s_waitcnt vmcnt(0)
+; GFX6-NEXT: v_and_b32_e32 v2, v2, v3
+; GFX6-NEXT: buffer_store_dword v2, v[0:1], s[0:3], 0 addr64
+; GFX6-NEXT: s_endpgm
+;
+; GFX8-LABEL: v_and_i32_vgpr_vgpr:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8-NEXT: s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8-NEXT: v_lshlrev_b32_e32 v4, 2, v0
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: v_mov_b32_e32 v1, s3
+; GFX8-NEXT: v_add_u32_e32 v0, vcc, s2, v4
+; GFX8-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
+; GFX8-NEXT: v_mov_b32_e32 v3, s5
+; GFX8-NEXT: v_add_u32_e32 v2, vcc, s4, v4
+; GFX8-NEXT: v_addc_u32_e32 v3, vcc, 0, v3, vcc
+; GFX8-NEXT: flat_load_dword v5, v[0:1]
+; GFX8-NEXT: flat_load_dword v2, v[2:3]
+; GFX8-NEXT: v_mov_b32_e32 v1, s1
+; GFX8-NEXT: v_add_u32_e32 v0, vcc, s0, v4
+; GFX8-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
+; GFX8-NEXT: s_waitcnt vmcnt(0)
+; GFX8-NEXT: v_and_b32_e32 v2, v5, v2
+; GFX8-NEXT: flat_store_dword v[0:1], v2
+; GFX8-NEXT: s_endpgm
+;
+; EG-LABEL: v_and_i32_vgpr_vgpr:
+; EG: ; %bb.0:
+; EG-NEXT: ALU 3, @10, KC0[CB0:0-32], KC1[]
+; EG-NEXT: TEX 1 @6
+; EG-NEXT: ALU 3, @14, KC0[CB0:0-32], KC1[]
+; EG-NEXT: MEM_RAT_CACHELESS STORE_RAW T0.X, T1.X, 1
+; EG-NEXT: CF_END
+; EG-NEXT: PAD
+; EG-NEXT: Fetch clause starting at 6:
+; EG-NEXT: VTX_READ_32 T1.X, T1.X, 0, #1
+; EG-NEXT: VTX_READ_32 T0.X, T0.X, 0, #1
+; EG-NEXT: ALU clause starting at 10:
+; EG-NEXT: LSHL * T0.W, T0.X, literal.x,
+; EG-NEXT: 2(2.802597e-45), 0(0.000000e+00)
+; EG-NEXT: ADD_INT T0.X, KC0[2].Z, PV.W,
+; EG-NEXT: ADD_INT * T1.X, KC0[2].W, PV.W,
+; EG-NEXT: ALU clause starting at 14:
+; EG-NEXT: AND_INT T0.X, T0.X, T1.X,
+; EG-NEXT: ADD_INT * T0.W, KC0[2].Y, T0.W,
+; EG-NEXT: LSHR * T1.X, PV.W, literal.x,
+; EG-NEXT: 2(2.802597e-45), 0(0.000000e+00)
%tid = call i32 @llvm.amdgcn.workitem.id.x() #0
%gep.a = getelementptr i32, ptr addrspace(1) %aptr, i32 %tid
%gep.b = getelementptr i32, ptr addrspace(1) %bptr, i32 %tid
@@ -105,11 +404,62 @@ define amdgpu_kernel void @v_and_i32_vgpr_vgpr(ptr addrspace(1) %out, ptr addrsp
ret void
}
-; FUNC-LABEL: {{^}}v_and_i32_sgpr_vgpr:
-; SI-DAG: s_load_dword [[SA:s[0-9]+]]
-; SI-DAG: {{buffer|flat}}_load_dword [[VB:v[0-9]+]]
-; SI: v_and_b32_e32 v{{[0-9]+}}, [[SA]], [[VB]]
define amdgpu_kernel void @v_and_i32_sgpr_vgpr(ptr addrspace(1) %out, i32 %a, ptr addrspace(1) %bptr) {
+; GFX6-LABEL: v_and_i32_sgpr_vgpr:
+; GFX6: ; %bb.0:
+; GFX6-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0xd
+; GFX6-NEXT: s_mov_b32 s3, 0xf000
+; GFX6-NEXT: s_mov_b32 s2, 0
+; GFX6-NEXT: v_lshlrev_b32_e32 v0, 2, v0
+; GFX6-NEXT: v_mov_b32_e32 v1, 0
+; GFX6-NEXT: s_waitcnt lgkmcnt(0)
+; GFX6-NEXT: buffer_load_dword v2, v[0:1], s[0:3], 0 addr64
+; GFX6-NEXT: s_load_dword s6, s[4:5], 0xb
+; GFX6-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x9
+; GFX6-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX6-NEXT: v_and_b32_e32 v2, s6, v2
+; GFX6-NEXT: buffer_store_dword v2, v[0:1], s[0:3], 0 addr64
+; GFX6-NEXT: s_endpgm
+;
+; GFX8-LABEL: v_and_i32_sgpr_vgpr:
+; GFX8: ; %bb.0:
+; GFX8-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x34
+; GFX8-NEXT: v_lshlrev_b32_e32 v2, 2, v0
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: v_mov_b32_e32 v1, s1
+; GFX8-NEXT: v_add_u32_e32 v0, vcc, s0, v2
+; GFX8-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
+; GFX8-NEXT: flat_load_dword v3, v[0:1]
+; GFX8-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8-NEXT: s_load_dword s2, s[4:5], 0x2c
+; GFX8-NEXT: s_waitcnt lgkmcnt(0)
+; GFX8-NEXT: v_mov_b32_e32 v1, s1
+; GFX8-NEXT: v_add_u32_e32 v0, vcc, s0, v2
+; GFX8-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
+; GFX8-NEXT: s_waitcnt vmcnt(0)
+; GFX8-NEXT: v_and_b32_e32 v2, s2, v3
+; GFX8-NEXT: flat_store_dword v[0:1], v2
+; GFX8-NEXT: s_endpgm
+;
+; EG-LABEL: v_and_i32_sgpr_vgpr:
+; EG: ; %bb.0:
+; EG-NEXT: ALU 2, @8, KC0[CB0:0-32], KC1[]
+; EG-NEXT: TEX 0 @6
+; EG-NEXT: ALU 3, @11, KC0[CB0:0-32], KC1[]
+; EG-NEXT: MEM_RAT_CACHELESS STORE_RAW T0.X, T1.X, 1
+; EG-NEXT: CF_END
+; EG-NEXT: PAD
+; EG-NEXT: Fetch clause starting at 6:
+; EG-NEXT: VTX_READ_32 T0.X, T0.X, 0, #1
+; EG-NEXT: ALU clause starting at 8:
+; EG-NEXT: LSHL * T0.W, T0.X, literal.x,
+; EG-NEXT: 2(2.802597e-45), 0(0.000000e+00)
+; EG-NEXT: ADD_INT * T0.X, KC0[2].W, PV.W,
+; EG-NEXT: ALU clause starting at 11:
+; EG-NEXT: AND_INT T0.X, KC0[2].Z, T0.X,
+; EG-NEXT: ADD_INT * T0.W, KC0[2].Y, T0.W,
+; EG-NEXT: LSHR * T1.X, PV.W, literal.x,
+; EG-NEXT: 2(2.802597e-45), 0(0.000000e+00)
%tid = call i32 @llvm.amdgcn.workitem.id.x() #0
%gep.b = getelementptr i32, ptr addrspace(1) %bptr, i32 %tid
%gep.out = getelementptr i32, ptr addrspace(1) %out, i32 %tid
@@ -119,11 +469,61 @@ define amdgpu_kernel void @v_and_i32_sgpr_vgpr(ptr addrspace(1) %out, i32 %a, pt
ret void
}
-; FUNC-LABEL: {{^}}v_and_i3...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/145371
More information about the llvm-commits
mailing list