| Issue |
119852
|
| Summary |
[mlir] Inconsistent output when executing MLIR program with `--linalg-fuse-elementwise-ops `
|
| Labels |
mlir
|
| Assignees |
|
| Reporter |
Emilyaxe
|
git version: 953838dceaf
system: `Ubuntu 18.04.6 LTS`
## Description:
I am experiencing an inconsistent result when executing the same MLIR program with and without `--linalg-fuse-elementwise-ops `.
## Steps to Reproduce:
### 1. **MLIR Program (a.mlir)**:
a.mlir:
```
module {
func.func private @printMemrefI32(tensor<*xi32>)
func.func private @printMemrefF32(tensor<*xf32>)
func.func @main() {
%1 = "tosa.const"() <{value = dense<5319> : tensor<1x28x1xi32>}> : () -> tensor<1x28x1xi32>
%2 = "tosa.const"() <{value = dense<-5633> : tensor<1x28x1xi32>}> : () -> tensor<1x28x1xi32>
%8 = tosa.int_div %1, %2 : (tensor<1x28x1xi32>, tensor<1x28x1xi32>) -> tensor<1x28x1xi32>
%9 = tosa.clz %8 : (tensor<1x28x1xi32>) -> tensor<1x28x1xi32>
%12 = tosa.logical_left_shift %8, %9 : (tensor<1x28x1xi32>, tensor<1x28x1xi32>) -> tensor<1x28x1xi32>
%cast = tensor.cast %12 : tensor<1x28x1xi32> to tensor<*xi32>
call @printMemrefI32(%cast) : (tensor<*xi32>) -> ()
return
}
}
```
### 2. **Command to Run without `--linalg-fuse-elementwise-ops ` :**
```
/data/szy/MLIR/llvm-release/llvm-project/build/bin/mlir-opt a.mlir -pass-pipeline="builtin.module(func.func(tosa-to-linalg-named,tosa-to-linalg))" | /data/szy/MLIR/llvm-release/llvm-project/build/bin/mlir-opt -tosa-to-arith -convert-linalg-to-loops -one-shot-bufferize="bufferize-function-boundaries" --expand-strided-metadata -convert-arith-to-llvm -test-math-polynomial-approximation -convert-arith-to-llvm -convert-math-to-llvm -convert-linalg-to-parallel-loops -finalize-memref-to-llvm -convert-scf-to-cf -convert-func-to-llvm -reconcile-unrealized-casts | timeout 10 /data/szy/MLIR/llvm-release/llvm-project/build/bin/mlir-cpu-runner -e main -entry-point-result=void --shared-libs=/data/szy/MLIR/llvm-release/llvm-project/build/lib/libmlir_c_runner_utils.so --shared-libs=/data/szy/MLIR/llvm-release/llvm-project/build/lib/libmlir_runner_utils.so --shared-libs=/data/szy/MLIR/llvm-release/llvm-project/build/lib/libmlir_async_runtime.so
```
### 3. **Output without `--linalg-fuse-elementwise-ops ` :**:
```
[[[0],
[0],
[0],
[0],
[0]]]
```
### 4. **Command to Run with `--linalg-fuse-elementwise-ops ` :**
```
/data/szy/MLIR/llvm-release/llvm-project/build/bin/mlir-opt a.mlir -pass-pipeline="builtin.module(func.func(tosa-to-linalg-named,tosa-to-linalg))" | /data/szy/MLIR/llvm-release/llvm-project/build/bin/mlir-opt -tosa-to-arith --linalg-fuse-elementwise-ops -convert-linalg-to-loops -one-shot-bufferize="bufferize-function-boundaries" --expand-strided-metadata -convert-arith-to-llvm -test-math-polynomial-approximation -convert-arith-to-llvm -convert-math-to-llvm -convert-linalg-to-parallel-loops -finalize-memref-to-llvm -convert-scf-to-cf -convert-func-to-llvm -reconcile-unrealized-casts | timeout 10 /data/szy/MLIR/llvm-release/llvm-project/build/bin/mlir-cpu-runner -e main -entry-point-result=void --shared-libs=/data/szy/MLIR/llvm-release/llvm-project/build/lib/libmlir_c_runner_utils.so --shared-libs=/data/szy/MLIR/llvm-release/llvm-project/build/lib/libmlir_runner_utils.so --shared-libs=/data/szy/MLIR/llvm-release/llvm-project/build/lib/libmlir_async_runtime.so
```
### 5. **Output with `--linalg-fuse-elementwise-ops ` :**
```
[[[64],
[22050],
[1],
[0],
[0]]]
```
![]()
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
