[llvm] 53becf5 - [AMDGPU] Regenerate shift-and-i128-ubfe.ll test checks
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Thu Nov 4 07:27:38 PDT 2021
Author: Simon Pilgrim
Date: 2021-11-04T14:27:30Z
New Revision: 53becf5df2d45fe9c65986d68f22fb78e15ef4b5
URL: https://github.com/llvm/llvm-project/commit/53becf5df2d45fe9c65986d68f22fb78e15ef4b5
DIFF: https://github.com/llvm/llvm-project/commit/53becf5df2d45fe9c65986d68f22fb78e15ef4b5.diff
LOG: [AMDGPU] Regenerate shift-and-i128-ubfe.ll test checks
Added:
Modified:
llvm/test/CodeGen/AMDGPU/shift-and-i128-ubfe.ll
Removed:
################################################################################
diff --git a/llvm/test/CodeGen/AMDGPU/shift-and-i128-ubfe.ll b/llvm/test/CodeGen/AMDGPU/shift-and-i128-ubfe.ll
index 36e02000608da..9f79364cf40f6 100644
--- a/llvm/test/CodeGen/AMDGPU/shift-and-i128-ubfe.ll
+++ b/llvm/test/CodeGen/AMDGPU/shift-and-i128-ubfe.ll
@@ -1,17 +1,28 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
; Extract the high bit of the 1st quarter
-; GCN-LABEL: {{^}}v_uextract_bit_31_i128:
-; GCN: buffer_load_dword [[VAL:v[0-9]+]], v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64{{$}}
-
-; GCN: v_mov_b32_e32 v[[ZERO0:[0-9]+]], 0{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO1:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO2:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN: v_lshrrev_b32_e32 v[[SHIFT:[0-9]+]], 31, [[VAL]]
-
-; GCN: buffer_store_dwordx4 v{{\[}}[[SHIFT]]:[[ZERO2]]{{\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64{{$}}
-; GCN: s_endpgm
define amdgpu_kernel void @v_uextract_bit_31_i128(i128 addrspace(1)* %out, i128 addrspace(1)* %in) #1 {
+; GCN-LABEL: v_uextract_bit_31_i128:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_load_dwordx4 s[8:11], s[0:1], 0x9
+; GCN-NEXT: s_ashr_i32 s3, s2, 31
+; GCN-NEXT: s_lshl_b64 s[0:1], s[2:3], 4
+; GCN-NEXT: v_mov_b32_e32 v5, s1
+; GCN-NEXT: s_mov_b32 s7, 0xf000
+; GCN-NEXT: s_mov_b32 s6, 0
+; GCN-NEXT: s_waitcnt lgkmcnt(0)
+; GCN-NEXT: s_mov_b64 s[4:5], s[10:11]
+; GCN-NEXT: v_mov_b32_e32 v4, s0
+; GCN-NEXT: buffer_load_dword v0, v[4:5], s[4:7], 0 addr64
+; GCN-NEXT: v_mov_b32_e32 v1, 0
+; GCN-NEXT: v_mov_b32_e32 v2, v1
+; GCN-NEXT: s_mov_b64 s[10:11], s[6:7]
+; GCN-NEXT: v_mov_b32_e32 v3, v1
+; GCN-NEXT: s_waitcnt vmcnt(0)
+; GCN-NEXT: v_lshrrev_b32_e32 v0, 31, v0
+; GCN-NEXT: buffer_store_dwordx4 v[0:3], v[4:5], s[8:11], 0 addr64
+; GCN-NEXT: s_endpgm
%id.x = tail call i32 @llvm.amdgcn.workgroup.id.x()
%in.gep = getelementptr i128, i128 addrspace(1)* %in, i32 %id.x
%out.gep = getelementptr i128, i128 addrspace(1)* %out, i32 %id.x
@@ -23,18 +34,26 @@ define amdgpu_kernel void @v_uextract_bit_31_i128(i128 addrspace(1)* %out, i128
}
; Extract the high bit of the 2nd quarter
-; GCN-LABEL: {{^}}v_uextract_bit_63_i128:
-; GCN-DAG: buffer_load_dword [[VAL:v[0-9]+]], v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64 offset:4{{$}}
-
-; GCN-DAG: v_mov_b32_e32 v[[ZERO0:[0-9]+]], 0{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO1:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO2:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO3:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN-DAG: v_lshrrev_b32_e32 v[[SHIFT:[0-9]+]], 31, [[VAL]]
-
-; GCN-DAG: buffer_store_dwordx4 v{{\[}}[[SHIFT]]:[[ZERO3]]{{\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64{{$}}
-; GCN: s_endpgm
define amdgpu_kernel void @v_uextract_bit_63_i128(i128 addrspace(1)* %out, i128 addrspace(1)* %in) #1 {
+; GCN-LABEL: v_uextract_bit_63_i128:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x9
+; GCN-NEXT: s_mov_b32 s7, 0xf000
+; GCN-NEXT: s_mov_b32 s6, 0
+; GCN-NEXT: v_lshlrev_b32_e32 v4, 4, v0
+; GCN-NEXT: v_mov_b32_e32 v5, 0
+; GCN-NEXT: s_waitcnt lgkmcnt(0)
+; GCN-NEXT: s_mov_b64 s[8:9], s[2:3]
+; GCN-NEXT: s_mov_b64 s[10:11], s[6:7]
+; GCN-NEXT: buffer_load_dword v0, v[4:5], s[8:11], 0 addr64 offset:4
+; GCN-NEXT: v_mov_b32_e32 v1, v5
+; GCN-NEXT: v_mov_b32_e32 v2, v5
+; GCN-NEXT: s_mov_b64 s[4:5], s[0:1]
+; GCN-NEXT: v_mov_b32_e32 v3, v5
+; GCN-NEXT: s_waitcnt vmcnt(0)
+; GCN-NEXT: v_lshrrev_b32_e32 v0, 31, v0
+; GCN-NEXT: buffer_store_dwordx4 v[0:3], v[4:5], s[4:7], 0 addr64
+; GCN-NEXT: s_endpgm
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%in.gep = getelementptr i128, i128 addrspace(1)* %in, i32 %id.x
%out.gep = getelementptr i128, i128 addrspace(1)* %out, i32 %id.x
@@ -46,17 +65,28 @@ define amdgpu_kernel void @v_uextract_bit_63_i128(i128 addrspace(1)* %out, i128
}
; Extract the high bit of the 3rd quarter
-; GCN-LABEL: {{^}}v_uextract_bit_95_i128:
-; GCN: buffer_load_dword [[VAL:v[0-9]+]], v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64 offset:8{{$}}
-
-; GCN-DAG: v_mov_b32_e32 v[[ZERO0:[0-9]+]], 0{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO1:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO2:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN-DAG: v_lshrrev_b32_e32 v[[SHIFT:[0-9]+]], 31, [[VAL]]
-
-; GCN-DAG: buffer_store_dwordx4 v{{\[}}[[SHIFT]]:[[ZERO2]]{{\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64{{$}}
-; GCN: s_endpgm
define amdgpu_kernel void @v_uextract_bit_95_i128(i128 addrspace(1)* %out, i128 addrspace(1)* %in) #1 {
+; GCN-LABEL: v_uextract_bit_95_i128:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_load_dwordx4 s[4:7], s[0:1], 0x9
+; GCN-NEXT: s_ashr_i32 s3, s2, 31
+; GCN-NEXT: s_lshl_b64 s[0:1], s[2:3], 4
+; GCN-NEXT: s_mov_b32 s3, 0xf000
+; GCN-NEXT: s_mov_b32 s2, 0
+; GCN-NEXT: v_mov_b32_e32 v5, s1
+; GCN-NEXT: s_waitcnt lgkmcnt(0)
+; GCN-NEXT: s_mov_b64 s[8:9], s[6:7]
+; GCN-NEXT: s_mov_b64 s[10:11], s[2:3]
+; GCN-NEXT: v_mov_b32_e32 v4, s0
+; GCN-NEXT: buffer_load_dword v0, v[4:5], s[8:11], 0 addr64 offset:8
+; GCN-NEXT: v_mov_b32_e32 v1, 0
+; GCN-NEXT: v_mov_b32_e32 v2, v1
+; GCN-NEXT: s_mov_b64 s[0:1], s[4:5]
+; GCN-NEXT: v_mov_b32_e32 v3, v1
+; GCN-NEXT: s_waitcnt vmcnt(0)
+; GCN-NEXT: v_lshrrev_b32_e32 v0, 31, v0
+; GCN-NEXT: buffer_store_dwordx4 v[0:3], v[4:5], s[0:3], 0 addr64
+; GCN-NEXT: s_endpgm
%id.x = tail call i32 @llvm.amdgcn.workgroup.id.x()
%in.gep = getelementptr i128, i128 addrspace(1)* %in, i32 %id.x
%out.gep = getelementptr i128, i128 addrspace(1)* %out, i32 %id.x
@@ -68,18 +98,26 @@ define amdgpu_kernel void @v_uextract_bit_95_i128(i128 addrspace(1)* %out, i128
}
; Extract the high bit of the 4th quarter
-; GCN-LABEL: {{^}}v_uextract_bit_127_i128:
-; GCN-DAG: buffer_load_dword [[VAL:v[0-9]+]], v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64 offset:12{{$}}
-
-; GCN-DAG: v_mov_b32_e32 v[[ZERO0:[0-9]+]], 0{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO1:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO2:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN: v_mov_b32_e32 v[[ZERO3:[0-9]+]], v[[ZERO0]]{{$}}
-; GCN-DAG: v_lshrrev_b32_e32 v[[SHIFT:[0-9]+]], 31, [[VAL]]
-
-; GCN-DAG: buffer_store_dwordx4 v{{\[}}[[SHIFT]]:[[ZERO3]]{{\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64{{$}}
-; GCN: s_endpgm
define amdgpu_kernel void @v_uextract_bit_127_i128(i128 addrspace(1)* %out, i128 addrspace(1)* %in) #1 {
+; GCN-LABEL: v_uextract_bit_127_i128:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x9
+; GCN-NEXT: s_mov_b32 s7, 0xf000
+; GCN-NEXT: s_mov_b32 s6, 0
+; GCN-NEXT: v_lshlrev_b32_e32 v4, 4, v0
+; GCN-NEXT: v_mov_b32_e32 v5, 0
+; GCN-NEXT: s_waitcnt lgkmcnt(0)
+; GCN-NEXT: s_mov_b64 s[8:9], s[2:3]
+; GCN-NEXT: s_mov_b64 s[10:11], s[6:7]
+; GCN-NEXT: buffer_load_dword v0, v[4:5], s[8:11], 0 addr64 offset:12
+; GCN-NEXT: v_mov_b32_e32 v1, v5
+; GCN-NEXT: v_mov_b32_e32 v2, v5
+; GCN-NEXT: s_mov_b64 s[4:5], s[0:1]
+; GCN-NEXT: v_mov_b32_e32 v3, v5
+; GCN-NEXT: s_waitcnt vmcnt(0)
+; GCN-NEXT: v_lshrrev_b32_e32 v0, 31, v0
+; GCN-NEXT: buffer_store_dwordx4 v[0:3], v[4:5], s[4:7], 0 addr64
+; GCN-NEXT: s_endpgm
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%in.gep = getelementptr i128, i128 addrspace(1)* %in, i32 %id.x
%out.gep = getelementptr i128, i128 addrspace(1)* %out, i32 %id.x
@@ -91,19 +129,26 @@ define amdgpu_kernel void @v_uextract_bit_127_i128(i128 addrspace(1)* %out, i128
}
; Spans more than 2 dword boundaries
-; GCN-LABEL: {{^}}v_uextract_bit_34_100_i128:
-; GCN-DAG: buffer_load_dwordx4 v{{\[}}[[VAL0:[0-9]+]]:[[VAL3:[0-9]+]]{{\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64{{$}}
-
-; GCN-DAG: v_lshl_b64 v{{\[}}[[SHLLO:[0-9]+]]:[[SHLHI:[0-9]+]]{{\]}}, v{{\[[0-9]+:[0-9]+\]}}, 30
-; GCN-DAG: v_lshrrev_b32_e32 v[[ELT1PART:[0-9]+]], 2, v{{[0-9]+}}
-; GCN-DAG: v_bfe_u32 v[[ELT2PART:[0-9]+]], v[[VAL3]], 2, 2{{$}}
-; GCN-DAG: v_mov_b32_e32 v[[ZERO:[0-9]+]], 0{{$}}
-; GCN-DAG: v_or_b32_e32 v[[OR0:[0-9]+]], v[[ELT1PART]], v[[SHLLO]]
-; GCN-DAG: v_mov_b32_e32 v[[ZERO1:[0-9]+]], v[[ZERO]]{{$}}
-
-; GCN-DAG: buffer_store_dwordx4 v{{\[}}[[OR0]]:[[ZERO1]]{{\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64{{$}}
-; GCN: s_endpgm
define amdgpu_kernel void @v_uextract_bit_34_100_i128(i128 addrspace(1)* %out, i128 addrspace(1)* %in) #1 {
+; GCN-LABEL: v_uextract_bit_34_100_i128:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_load_dwordx4 s[4:7], s[0:1], 0x9
+; GCN-NEXT: s_mov_b32 s3, 0xf000
+; GCN-NEXT: s_mov_b32 s2, 0
+; GCN-NEXT: v_lshlrev_b32_e32 v8, 4, v0
+; GCN-NEXT: v_mov_b32_e32 v9, 0
+; GCN-NEXT: s_waitcnt lgkmcnt(0)
+; GCN-NEXT: s_mov_b64 s[0:1], s[6:7]
+; GCN-NEXT: buffer_load_dwordx4 v[0:3], v[8:9], s[0:3], 0 addr64
+; GCN-NEXT: s_mov_b64 s[6:7], s[2:3]
+; GCN-NEXT: v_mov_b32_e32 v7, v9
+; GCN-NEXT: s_waitcnt vmcnt(0)
+; GCN-NEXT: v_lshl_b64 v[4:5], v[2:3], 30
+; GCN-NEXT: v_lshrrev_b32_e32 v0, 2, v1
+; GCN-NEXT: v_bfe_u32 v6, v3, 2, 2
+; GCN-NEXT: v_or_b32_e32 v4, v0, v4
+; GCN-NEXT: buffer_store_dwordx4 v[4:7], v[8:9], s[4:7], 0 addr64
+; GCN-NEXT: s_endpgm
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%in.gep = getelementptr i128, i128 addrspace(1)* %in, i32 %id.x
%out.gep = getelementptr i128, i128 addrspace(1)* %out, i32 %id.x
More information about the llvm-commits
mailing list