[Mlir-commits] [mlir] [mlir][complex] Make CPU runner test platform agnostic (PR #85607)

[llvmlistbot at llvm.org]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-mlir

Author: Kai Sasaki (Lewuathe)

<details>
<summary>Changes</summary>

The `correctness.mlir` test for complex dialects has been excluding the AArch64 platform. We have found that this is mainly due to the uncommon treatment of the Apple Silicon platform, which does not recognize signed NaN. We did not this behavior in AArch64 platform in general. Therefore, it is reasonable to make the test platform agnostic instead of keeping this test from running in the AArch64 platform, considering that the signness check of NaN value is not so essential here.

I confirmed this test passed in my Apple Silicon environment here.

```
$ uname -a
Darwin Kais-Mac.local 23.3.0 Darwin Kernel Version 23.3.0: Wed Dec 20 21:30:44 PST 2023; root:xnu-10002.81.5~7/RELEASE_ARM64_T6000 arm64
```

See more detail: https://github.com/llvm/llvm-project/issues/58531



---
Full diff: https://github.com/llvm/llvm-project/pull/85607.diff


1 Files Affected:

- (modified) mlir/test/Integration/Dialect/Complex/CPU/correctness.mlir (+2-5) 


``````````diff
diff --git a/mlir/test/Integration/Dialect/Complex/CPU/correctness.mlir b/mlir/test/Integration/Dialect/Complex/CPU/correctness.mlir
index a42ed6968d3700..4a334b73baa25e 100644
--- a/mlir/test/Integration/Dialect/Complex/CPU/correctness.mlir
+++ b/mlir/test/Integration/Dialect/Complex/CPU/correctness.mlir
@@ -9,9 +9,6 @@
 // RUN:  -shared-libs=%mlir_c_runner_utils |\
 // RUN: FileCheck %s
 
-// XFAIL: target=aarch64{{.*}}
-// See: https://github.com/llvm/llvm-project/issues/58531
-
 func.func @test_unary(%input: tensor<?xcomplex<f32>>,
                       %func: (complex<f32>) -> complex<f32>) {
   %c0 = arith.constant 0 : index
@@ -189,8 +186,8 @@ func.func @entry() {
     // CHECK-NEXT:  0
     // CHECK-NEXT:  0
     (0.0, 0.0), (-1.0, 0.0),
-    // CHECK-NEXT:  -nan
-    // CHECK-NEXT:  -nan
+    // CHECK-NEXT:  nan
+    // CHECK-NEXT:  nan
     (1.0, 1.0), (1.0, 1.0)
     // CHECK-NEXT:  0.273
     // CHECK-NEXT:  0.583

``````````

</details>


https://github.com/llvm/llvm-project/pull/85607