[Mlir-commits] [mlir] ea28668 - [mlir][ArmSME] Remove XFAILs (#104758)

[llvmlistbot at llvm.org]

Author: Benjamin Maxwell
Date: 2024-08-19T14:55:09+01:00
New Revision: ea28668f8c980c9987b0e2f305b0a8eff9797401

URL: https://github.com/llvm/llvm-project/commit/ea28668f8c980c9987b0e2f305b0a8eff9797401
DIFF: https://github.com/llvm/llvm-project/commit/ea28668f8c980c9987b0e2f305b0a8eff9797401.diff

LOG: [mlir][ArmSME] Remove XFAILs (#104758)

Buildbots have been updated to QEMU 9.0.2: 
https://github.com/llvm/llvm-project/pull/104758#issuecomment-2296592845

Added: 
    

Modified: 
    mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/multi-tile-matmul-mixed-types.mlir
    mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-i8i8i32.mlir

Removed: 
    


################################################################################
diff  --git a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/multi-tile-matmul-mixed-types.mlir b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/multi-tile-matmul-mixed-types.mlir
index 8a4c8688d64748..78815a38612e96 100644
--- a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/multi-tile-matmul-mixed-types.mlir
+++ b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/multi-tile-matmul-mixed-types.mlir
@@ -13,14 +13,8 @@
 /// tests a mixed i8 to i32 matmul and outer product fusion which fuses 16
 /// outer products (four per tile) into four 4-way outer products.
 
-/// NOTE: QEMU gives incorrect result for SME SMOPA 4-way outer product
-/// instruction (version <= 8.2.0, latest version at time of writing), see:
-/// https://gitlab.com/qemu-project/qemu/-/issues/2083
-/// This test is expected to fail until a fixed version of QEMU can be used.
-
-/// FIXME: Remove the 'XFAIL' below once a fixed QEMU version is available
-/// (and installed on CI buildbot).
-/// XFAIL: *
+// NOTE: QEMU <= 8.2.0 gives incorrect result for SME SMOPA 4-way outer product
+// instruction see: https://gitlab.com/qemu-project/qemu/-/issues/2083.
 
 func.func @matmul_i8_to_i32(%A : tensor<?x?xi8>, %B : tensor<?x?xi8>, %C : tensor<?x?xi32>) {
   %res = linalg.matmul ins(%A, %B: tensor<?x?xi8>, tensor<?x?xi8>)

diff  --git a/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-i8i8i32.mlir b/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-i8i8i32.mlir
index 653b0d7beb2c78..74ddbc63380b70 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-i8i8i32.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-i8i8i32.mlir
@@ -7,14 +7,8 @@
 
 // RUN: %{compile} | %{run} | FileCheck %s
 
-// NOTE: QEMU gives incorrect result for SME SMOPA 4-way outer product
-// instruction (version <= 8.2.0, latest version at time of writing), see:
-// https://gitlab.com/qemu-project/qemu/-/issues/2083 This test is expected to
-// fail (CHECK lines are correct) until a fixed version of QEMU can be used.
-
-// FIXME: Remove the 'XFAIL' below once a fixed QEMU version is available
-// (and installed on CI buildbot).
-// XFAIL: *
+// NOTE: QEMU <= 8.2.0 gives incorrect result for SME SMOPA 4-way outer product
+// instruction see: https://gitlab.com/qemu-project/qemu/-/issues/2083.
 
 // NOTE: there is no non-widening variant for these types and this test can't
 // be lowered without the widening pass, therefore we can't check if the result