[Mlir-commits] [mlir] [MLIR][XeGPU] Extend Wg-to-Sg Distribution of Multi-Reduction Op for round-robin layout (PR #189988)
llvmlistbot at llvm.org
llvmlistbot at llvm.org
Wed Apr 1 09:12:17 PDT 2026
github-actions[bot] wrote:
<!--PREMERGE ADVISOR COMMENT: Windows-->
# :window: Windows x64 Test Results
* 3597 tests passed
* 415 tests skipped
* 2 tests failed
## Failed Tests
(click on a test name to see its output)
### MLIR
<details>
<summary>MLIR.Dialect/XeGPU/xegpu-wg-to-sg-unify-ops-rr.mlir</summary>
```
Exit Code: 2
Command Output (stdout):
--
# RUN: at line 1
c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe --xegpu-wg-to-sg-distribute -split-input-file C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops-rr.mlir | c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops-rr.mlir
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe' --xegpu-wg-to-sg-distribute -split-input-file 'C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops-rr.mlir'
# .---command stderr------------
# | Assertion failed: Index < size() && "invalid index for value range", file C:\_work\llvm-project\llvm-project\llvm\include\llvm/ADT/STLExtras.h, line 1249
# | PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace and instructions to reproduce the bug.
# | Stack dump:
# | 0. Program arguments: c:\\_work\\llvm-project\\llvm-project\\build\\bin\\mlir-opt.exe --xegpu-wg-to-sg-distribute -split-input-file C:\\_work\\llvm-project\\llvm-project\\mlir\\test\\Dialect\\XeGPU\\xegpu-wg-to-sg-unify-ops-rr.mlir
# | Exception Code: 0xC000001D
# | #0 0x00007ff62281b3f6 mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x375b3f6)
# | #1 0x00007fff7460bb04 (C:\Windows\System32\ucrtbase.dll+0x7bb04)
# | #2 0x00007fff7460cad1 (C:\Windows\System32\ucrtbase.dll+0x7cad1)
# | #3 0x00007fff7460e4a1 (C:\Windows\System32\ucrtbase.dll+0x7e4a1)
# | #4 0x00007fff7460e6e1 (C:\Windows\System32\ucrtbase.dll+0x7e6e1)
# | #5 0x00007ff62197c39e mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x28bc39e)
# | #6 0x00007ff62191b71b mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x285b71b)
# | #7 0x00007ff6219987a4 mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x28d87a4)
# | #8 0x00007ff6230ba057 mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x3ffa057)
# | #9 0x00007ff6230b6b2f mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x3ff6b2f)
# | #10 0x00007ff62199a92d mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x28da92d)
# | #11 0x00007ff621999e9b mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x28d9e9b)
# | #12 0x00007ff62199b3e2 mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x28db3e2)
# | #13 0x00007ff6219adbc1 mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x28edbc1)
# | #14 0x00007ff6219a19b9 mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x28e19b9)
# | #15 0x00007ff6219a1af2 mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x28e1af2)
# | #16 0x00007ff621967f94 mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x28a7f94)
# | #17 0x00007ff61f455b87 (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x395b87)
# | #18 0x00007ff61f4564b6 (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x3964b6)
# | #19 0x00007ff61f45c47e (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x39c47e)
# | #20 0x00007ff61f45bf13 (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x39bf13)
# | #21 0x00007ff61f0d08e7 (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x108e7)
# | #22 0x00007ff61f0cfa9f (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0xfa9f)
# | #23 0x00007ff61f446290 (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x386290)
# | #24 0x00007ff61f445ea1 (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x385ea1)
# | #25 0x00007ff61f0c383e (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x383e)
# | #26 0x00007ff61f0c3bee (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x3bee)
# | #27 0x00007ff61f0c3fa4 (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x3fa4)
# | #28 0x00007ff61f0c1450 (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x1450)
# | #29 0x00007ff627228400 mlir::detail::FallbackTypeIDResolver::registerImplicitTypeID(class llvm::StringRef) (c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe+0x8168400)
# | #30 0x00007fff81994cb0 (C:\Windows\System32\KERNEL32.DLL+0x14cb0)
# | #31 0x00007fff867dedcb (C:\Windows\SYSTEM32\ntdll.dll+0x7edcb)
# `-----------------------------
# error: command failed with exit status: 0xc000001d
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe' 'C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops-rr.mlir'
# .---command stderr------------
# | FileCheck error: '<stdin>' is empty.
# | FileCheck command line: c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops-rr.mlir
# `-----------------------------
# error: command failed with exit status: 2
--
```
</details>
<details>
<summary>MLIR.Dialect/XeGPU/xegpu-wg-to-sg-unify-ops.mlir</summary>
```
Exit Code: 1
Command Output (stdout):
--
# RUN: at line 1
c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe --xegpu-wg-to-sg-distribute -split-input-file C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir | c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe' --xegpu-wg-to-sg-distribute -split-input-file 'C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir'
# note: command had no output on stdout or stderr
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe' 'C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir'
# .---command stderr------------
# | C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir:3:15: error: CHECK-DAG: expected string not found in input
# | // CHECK-DAG: #map = affine_map<()[s0] -> (s0 floordiv 4)>
# | ^
# | <stdin>:1:1: note: scanning from here
# | module {
# | ^
# | <stdin>:2:2: note: possible intended match here
# | gpu.module @test_distribution {
# | ^
# | C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir:684:16: error: CHECK-DAG: expected string not found in input
# | // CHECK-DAG: %[[AFF0:.*]] = affine.apply #map2()[%[[SGID]]]
# | ^
# | <stdin>:967:62: note: scanning from here
# | gpu.func @vector_reduce_cross_sg_dim_1(%arg0: memref<?xf32>) {
# | ^
# | <stdin>:967:62: note: with "SGID" equal to "4"
# | gpu.func @vector_reduce_cross_sg_dim_1(%arg0: memref<?xf32>) {
# | ^
# | <stdin>:980:2: note: possible intended match here
# | %c32 = arith.constant 32 : index
# | ^
# | C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir:728:16: error: CHECK-DAG: expected string not found in input
# | // CHECK-DAG: %[[AFFINE1:.*]] = affine.apply {{#map[[:alnum:]_]*}}()[%[[SGID2]]]
# | ^
# | <stdin>:1028:68: note: scanning from here
# | gpu.func @vector_reduce_cross_sg_dim_0(%arg0: memref<256x128xf32>) {
# | ^
# | <stdin>:1028:68: note: with "SGID2" equal to "13"
# | gpu.func @vector_reduce_cross_sg_dim_0(%arg0: memref<256x128xf32>) {
# | ^
# | <stdin>:1037:2: note: possible intended match here
# | %4 = arith.muli %3, %c32 : index
# | ^
# | C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir:764:16: error: CHECK-DAG: expected string not found in input
# | // CHECK-DAG: %[[AFFINE0:.*]] = affine.apply {{#map[[:alnum:]_]*}}()[%[[SGID]]]
# | ^
# | <stdin>:1094:57: note: scanning from here
# | gpu.func @vector_reduce_multi_dim(%arg0: memref<?xf32>) {
# | ^
# | <stdin>:1094:57: note: with "SGID" equal to "4"
# | gpu.func @vector_reduce_multi_dim(%arg0: memref<?xf32>) {
# | ^
# | <stdin>:1102:4: note: possible intended match here
# | %alloca = memref.alloca() : memref<256xi8, 3>
# | ^
# | C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir:814:16: error: CHECK-DAG: expected string not found in input
# | // CHECK-DAG: %[[AFFINE0:.*]] = affine.apply {{#map[[:alnum:]_]*}}()[%[[SGID]]]
# | ^
# | <stdin>:1171:82: note: scanning from here
# | gpu.func @vector_reduce_multi_dim_nou_unit_local_reduction(%arg0: memref<?xf32>) {
# | ^
# | <stdin>:1171:82: note: with "SGID" equal to "4"
# | gpu.func @vector_reduce_multi_dim_nou_unit_local_reduction(%arg0: memref<?xf32>) {
# | ^
# | <stdin>:1194:2: note: possible intended match here
# | %12 = arith.muli %11, %c16 : index
# | ^
# |
# | Input file: <stdin>
# | Check file: C:\_work\llvm-project\llvm-project\mlir\test\Dialect\XeGPU\xegpu-wg-to-sg-unify-ops.mlir
# |
# | -dump-input=help explains the following input dump.
# |
# | Input was:
# | <<<<<<
# | 1: module {
# | dag:3'0 X~~~~~~~~ error: no match found
# | 2: gpu.module @test_distribution {
# | dag:3'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | dag:3'1 ? possible intended match
# | 3: gpu.func @create_nd_tdesc_no_offset(%arg0: memref<256x128xf32>) {
# | dag:3'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 4: %0 = xegpu.create_nd_tdesc %arg0 : memref<256x128xf32> -> !xegpu.tensor_desc<32x32xf32, #xegpu.layout<lane_layout = [1, 16], lane_data = [1, 1]>>
# | 5: gpu.return
# | 6: }
# | 7: gpu.func @create_nd_tdesc_with_ptr(%arg0: ui64, %arg1: index, %arg2: index, %arg3: index, %arg4: index) {
# | .
# | .
# | .
# | 962: %cst_0 = arith.constant dense<true> : vector<32xi1>
# | 963: %0 = xegpu.load %alloca[%cst], %cst_0 <{chunk_size = 1 : i64, layout = #xegpu.slice<#xegpu.layout<inst_data = [8, 16]>, dims = [0]>}> : memref<4096xf32>, vector<32xindex>, vector<32xi1> -> vector<32xf32>
# | 964: %1 = vector.broadcast %0 : vector<32xf32> to vector<32x32xf32>
# | 965: gpu.return
# | 966: }
# | 967: gpu.func @vector_reduce_cross_sg_dim_1(%arg0: memref<?xf32>) {
# | dag:684'0 X~~ error: no match found
# | dag:684'1 with "SGID" equal to "4"
# | 968: %cst = arith.constant dense<1.000000e+00> : vector<1x32xf32>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 969: %cst_0 = arith.constant dense<0> : vector<1x1x32xindex>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 970: %cst_1 = arith.constant dense<true> : vector<1x1x32xi1>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 971: %0 = xegpu.load %arg0[%cst_0], %cst_1 <{chunk_size = 1 : i64}> : memref<?xf32>, vector<1x1x32xindex>, vector<1x1x32xi1> -> vector<1x1x32xf32>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 972: %cst_2 = arith.constant dense<0.000000e+00> : vector<1x32xf32>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | .
# | .
# | .
# | 975: %alloca = memref.alloca() : memref<4096xi8, 3>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 976: %3 = xegpu.create_mem_desc %alloca : memref<4096xi8, 3> -> !xegpu.mem_desc<1x32x32xf32>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 977: %4 = gpu.subgroup_id : index
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 978: %c1 = arith.constant 1 : index
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 979: %c0 = arith.constant 0 : index
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 980: %c32 = arith.constant 32 : index
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | dag:684'2 ? possible intended match
# | 981: %5 = arith.remui %4, %c32 : index
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 982: %6 = arith.divui %4, %c32 : index
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 983: %c1_3 = arith.constant 1 : index
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 984: %c0_4 = arith.constant 0 : index
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 985: %c1_5 = arith.constant 1 : index
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | .
# | .
# | .
# | 1023: %cst_34 = arith.constant dense<0.000000e+00> : vector<1x32xf32>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1024: %14 = vector.multi_reduction <add>, %13, %cst_34 [1] : vector<1x32x32xf32> to vector<1x32xf32>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1025: %15 = arith.addf %14, %cst : vector<1x32xf32>
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1026: gpu.return
# | dag:684'0 ~~~~~~~~~~~~
# | 1027: }
# | dag:684'0 ~~~
# | 1028: gpu.func @vector_reduce_cross_sg_dim_0(%arg0: memref<256x128xf32>) {
# | dag:684'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | dag:728'0 X~~ error: no match found
# | dag:728'1 with "SGID2" equal to "13"
# | 1029: %cst = arith.constant dense<0.000000e+00> : vector<32xf32>
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1030: %0 = gpu.subgroup_id : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1031: %c4 = arith.constant 4 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1032: %1 = arith.remui %0, %c4 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1033: %2 = arith.divui %0, %c4 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1034: %c8 = arith.constant 8 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1035: %3 = arith.remui %2, %c8 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1036: %c32 = arith.constant 32 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1037: %4 = arith.muli %3, %c32 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | dag:728'2 ? possible intended match
# | 1038: %c32_0 = arith.constant 32 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1039: %5 = arith.muli %1, %c32_0 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1040: %c0 = arith.constant 0 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1041: %c0_1 = arith.constant 0 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1042: %c256 = arith.constant 256 : index
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | .
# | .
# | .
# | 1089: %cst_21 = arith.constant dense<0.000000e+00> : vector<32xf32>
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1090: %29 = vector.multi_reduction <add>, %28, %cst_21 [0] : vector<8x32xf32> to vector<32xf32>
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1091: %30 = arith.addf %29, %cst : vector<32xf32>
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1092: gpu.return
# | dag:728'0 ~~~~~~~~~~~~
# | 1093: }
# | dag:728'0 ~~~
# | 1094: gpu.func @vector_reduce_multi_dim(%arg0: memref<?xf32>) {
# | dag:728'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | dag:764'0 X~~ error: no match found
# | dag:764'1 with "SGID" equal to "4"
# | 1095: %cst = arith.constant dense<0.000000e+00> : vector<1x1xf32>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1096: %cst_0 = arith.constant dense<0> : vector<1x1x32x32xindex>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1097: %cst_1 = arith.constant dense<true> : vector<1x1x32x32xi1>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1098: %0 = xegpu.load %arg0[%cst_0], %cst_1 <{chunk_size = 1 : i64}> : memref<?xf32>, vector<1x1x32x32xindex>, vector<1x1x32x32xi1> -> vector<1x1x32x32xf32>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1099: %cst_2 = arith.constant dense<0.000000e+00> : vector<1x1xf32>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1100: %1 = vector.multi_reduction <add>, %0, %cst_2 [2, 3] : vector<1x1x32x32xf32> to vector<1x1xf32>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1101: %2 = vector.shape_cast %1 : vector<1x1xf32> to vector<1x1x1x1xf32>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1102: %alloca = memref.alloca() : memref<256xi8, 3>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | dag:764'2 ? possible intended match
# | 1103: %3 = xegpu.create_mem_desc %alloca : memref<256xi8, 3> -> !xegpu.mem_desc<2x2x4x4xf32>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1104: %4 = gpu.subgroup_id : index
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1105: %c4 = arith.constant 4 : index
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1106: %5 = arith.remui %4, %c4 : index
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1107: %6 = arith.divui %4, %c4 : index
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | .
# | .
# | .
# | 1166: %cst_31 = arith.constant dense<0.000000e+00> : vector<1x1xf32>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1167: %32 = vector.multi_reduction <add>, %31, %cst_31 [2, 3] : vector<1x1x4x4xf32> to vector<1x1xf32>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1168: %33 = arith.addf %32, %cst : vector<1x1xf32>
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1169: gpu.return
# | dag:764'0 ~~~~~~~~~~~~
# | 1170: }
# | dag:764'0 ~~~
# | 1171: gpu.func @vector_reduce_multi_dim_nou_unit_local_reduction(%arg0: memref<?xf32>) {
# | dag:764'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | dag:814'0 X~~ error: no match found
# | dag:814'1 with "SGID" equal to "4"
# | 1172: %cst = arith.constant dense<0.000000e+00> : vector<16x16xf32>
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1173: %cst_0 = arith.constant dense<0> : vector<16x16x32x32xindex>
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1174: %cst_1 = arith.constant dense<true> : vector<16x16x32x32xi1>
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1175: %0 = xegpu.load %arg0[%cst_0], %cst_1 <{chunk_size = 1 : i64}> : memref<?xf32>, vector<16x16x32x32xindex>, vector<16x16x32x32xi1> -> vector<16x16x32x32xf32>
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1176: %cst_2 = arith.constant dense<0.000000e+00> : vector<16x16xf32>
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | .
# | .
# | .
# | 1189: %9 = arith.remui %8, %c2 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1190: %10 = arith.divui %8, %c2 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1191: %c2_4 = arith.constant 2 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1192: %11 = arith.remui %10, %c2_4 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1193: %c16 = arith.constant 16 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1194: %12 = arith.muli %11, %c16 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | dag:814'2 ? possible intended match
# | 1195: %c16_5 = arith.constant 16 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1196: %13 = arith.muli %9, %c16_5 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1197: %c4_6 = arith.constant 4 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1198: %14 = arith.muli %7, %c4_6 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | 1199: %c4_7 = arith.constant 4 : index
# | dag:814'0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | .
# | .
# | .
# | >>>>>>
# `-----------------------------
# error: command failed with exit status: 1
--
```
</details>
If these failures are unrelated to your changes (for example tests are broken or flaky at HEAD), please open an issue at https://github.com/llvm/llvm-project/issues and add the `infrastructure` label.
https://github.com/llvm/llvm-project/pull/189988
More information about the Mlir-commits
mailing list