[Mlir-commits] [mlir] [mlir][affine]make affine-loop-unroll a FunctionOpInterface pass. (PR #126475)
Kunwar Grover
llvmlistbot at llvm.org
Wed Feb 12 18:09:16 PST 2025
================
@@ -1,8 +1,8 @@
-// RUN: mlir-opt -allow-unregistered-dialect %s -affine-loop-unroll="unroll-full" | FileCheck %s --check-prefix UNROLL-FULL
-// RUN: mlir-opt -allow-unregistered-dialect %s -affine-loop-unroll="unroll-full unroll-full-threshold=2" | FileCheck %s --check-prefix SHORT
-// RUN: mlir-opt -allow-unregistered-dialect %s -affine-loop-unroll="unroll-factor=4" | FileCheck %s --check-prefix UNROLL-BY-4
-// RUN: mlir-opt -allow-unregistered-dialect %s -affine-loop-unroll="unroll-factor=1" | FileCheck %s --check-prefix UNROLL-BY-1
-// RUN: mlir-opt -allow-unregistered-dialect %s -affine-loop-unroll="unroll-factor=5 cleanup-unroll=true" | FileCheck %s --check-prefix UNROLL-CLEANUP-LOOP
+// RUN: mlir-opt -allow-unregistered-dialect %s -pass-pipeline="builtin.module(func.func(affine-loop-unroll{unroll-full=true}),gpu.module(gpu.func(affine-loop-unroll{unroll-full=true})))" | FileCheck %s --check-prefix UNROLL-FULL
+// RUN: mlir-opt -allow-unregistered-dialect %s -pass-pipeline="builtin.module(func.func(affine-loop-unroll{unroll-full=true unroll-full-threshold=2}),gpu.module(gpu.func(affine-loop-unroll{unroll-full=true unroll-full-threshold=2})))" | FileCheck %s --check-prefix SHORT
+// RUN: mlir-opt -allow-unregistered-dialect %s -pass-pipeline="builtin.module(func.func(affine-loop-unroll{unroll-factor=4}),gpu.module(gpu.func(affine-loop-unroll{unroll-factor=4})))" | FileCheck %s --check-prefix UNROLL-BY-4
+// RUN: mlir-opt -allow-unregistered-dialect %s -pass-pipeline="builtin.module(func.func(affine-loop-unroll{unroll-factor=1}),gpu.module(gpu.func(affine-loop-unroll{unroll-factor=1})))" | FileCheck %s --check-prefix UNROLL-BY-1
+// RUN: mlir-opt -allow-unregistered-dialect %s -pass-pipeline="builtin.module(func.func(affine-loop-unroll{unroll-factor=5 cleanup-unroll=true}),gpu.module(gpu.func(affine-loop-unroll{unroll-factor=5 cleanup-unroll=true})))" | FileCheck %s --check-prefix UNROLL-CLEANUP-LOOP
----------------
Groverkss wrote:
I don't know how good of an idea is to run a pass twice. You can just add another RUN: with a check-prefix for gpu.func and just add a single test to see if it works.
https://github.com/llvm/llvm-project/pull/126475
More information about the Mlir-commits
mailing list