[Mlir-commits] [mlir] [mlir][gpu] Support `minimumnumf` and `maximumnumf` in GPU dialect (PR #226903)

[llvmlistbot at llvm.org]

github-actions[bot] wrote:

<!--PREMERGE ADVISOR COMMENT: Windows-->
# :window: Windows x64 Test Results

* 4219 tests passed
* 459 tests skipped
* 1 test failed

## Failed Tests
(click on a test name to see its output)

### MLIR
<details>
<summary>MLIR.Dialect/GPU/subgroup-reduce-lowering.mlir</summary>

```
Exit Code: 1

Command Output (stdout):
--
# RUN: at line 1
c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe    --test-gpu-subgroup-reduce-lowering C:\_work\llvm-project\llvm-project\mlir\test\Dialect\GPU\subgroup-reduce-lowering.mlir    | c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe C:\_work\llvm-project\llvm-project\mlir\test\Dialect\GPU\subgroup-reduce-lowering.mlir --check-prefix=CHECK-SUB
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe' --test-gpu-subgroup-reduce-lowering 'C:\_work\llvm-project\llvm-project\mlir\test\Dialect\GPU\subgroup-reduce-lowering.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\GPU\subgroup-reduce-lowering.mlir' --check-prefix=CHECK-SUB
# note: command had no output on stdout or stderr
# RUN: at line 5
c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe    --test-gpu-subgroup-reduce-lowering="expand-to-shuffles" C:\_work\llvm-project\llvm-project\mlir\test\Dialect\GPU\subgroup-reduce-lowering.mlir    | c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe C:\_work\llvm-project\llvm-project\mlir\test\Dialect\GPU\subgroup-reduce-lowering.mlir --check-prefix=CHECK-SHFL
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\mlir-opt.exe' --test-gpu-subgroup-reduce-lowering=expand-to-shuffles 'C:\_work\llvm-project\llvm-project\mlir\test\Dialect\GPU\subgroup-reduce-lowering.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\GPU\subgroup-reduce-lowering.mlir' --check-prefix=CHECK-SHFL
# .---command stderr------------
# | C:\_work\llvm-project\llvm-project\mlir\test\Dialect\GPU\subgroup-reduce-lowering.mlir:38:17: error: CHECK-SHFL: expected string not found in input
# |  // CHECK-SHFL: arith.maximumnumf
# |                 ^
# | <stdin>:23:26: note: scanning from here
# |  gpu.func @maximum_number(%arg0: vector<2xf32>) kernel {
# |                          ^
# | <stdin>:24:21: note: possible intended match here
# |  %0 = gpu.subgroup_reduce maximumnumf %arg0 : (vector<2xf32>) -> vector<2xf32>
# |                     ^
# | 
# | Input file: <stdin>
# | Check file: C:\_work\llvm-project\llvm-project\mlir\test\Dialect\GPU\subgroup-reduce-lowering.mlir
# | 
# | -dump-input=help explains the following input dump.
# | 
# | Input was:
# | <<<<<<
# |             .
# |             .
# |             .
# |            18:  %shuffleResult_6, %valid_7 = gpu.shuffle xor %3, %c16_i32, %c32_i32 : f32 
# |            19:  %4 = arith.minimumnumf %3, %shuffleResult_6 : f32 
# |            20:  "test.consume"(%4) : (f32) -> () 
# |            21:  gpu.return 
# |            22:  } 
# |            23:  gpu.func @maximum_number(%arg0: vector<2xf32>) kernel { 
# | check:38'0                             {                                  search range start (exclusive)
# | check:38'1                                                                error: no match found in search range
# |            24:  %0 = gpu.subgroup_reduce maximumnumf %arg0 : (vector<2xf32>) -> vector<2xf32> 
# | check:38'2                         ?                                                            possible intended match
# |            25:  "test.consume"(%0) : (vector<2xf32>) -> () 
# |            26:  gpu.return 
# |            27:  } 
# |            28:  gpu.func @kernel0(%arg0: vector<5xf16>) kernel { 
# | check:38'3                        }                                search range end (exclusive)
# |            29:  %c16_i32 = arith.constant 16 : i32 
# |            30:  %c8_i32 = arith.constant 8 : i32 
# |            31:  %c4_i32 = arith.constant 4 : i32 
# |            32:  %c2_i32 = arith.constant 2 : i32 
# |            33:  %c1_i32 = arith.constant 1 : i32 
# |             .
# |             .
# |             .
# | >>>>>>
# `-----------------------------
# 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/226903