| Issue |
149954
|
| Summary |
[MLIR]`tosa.exp` on large constant produces inf
|
| Labels |
mlir
|
| Assignees |
|
| Reporter |
sweead
|
test commit: [f3a3270](https://github.com/llvm/llvm-project/commit/f3a3270dbca3649b7d56aaa42cb8481fb34e2d67)
## Description:
When applying `tosa.exp` to a large constant value such as 187.84, the resulting value becomes `inf` due to floating-point overflow.
## Steps to Reproduce:
### Minimal MLIR program (test.mlir):
```
module {
func.func private @printMemrefF32(tensor<*xf32>)
func.func @main() {
%0 = "tosa.const"() <{values = dense<1.878400e+02> : tensor<f32>}> : () -> tensor<f32>
%1 = tosa.exp %0 : (tensor<f32>) -> tensor<f32>
%cast = tensor.cast %1 : tensor<f32> to tensor<*xf32>
call @printMemrefF32(%cast) : (tensor<*xf32>) -> ()
return
}
}
```
### Command:
```
mlir-opt test.mlir --tosa-to-linalg-pipeline --sparsifier | \
mlir-runner -e main -entry-point-result=void -shared-libs=/home/workdir/llvm-project/build/lib/libmlir_runner_utils.so
```
### Output:
```
[inf]
```
![]()
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
