pymlir
pymlir copied to clipboard
spcl
Python interface for MLIR - the Multi-Level Intermediate Representation
I haven't been able to reproduce the problem described in these comments: ``` # Pre-transform code to avoid parsing issues with ceildiv/floordiv/mod, # in which two symbols could be parsed...
- Added module `mlir.run` that contains helpers for calling MLIR functions. - Demonstrated some examples under `tests/test_run.py`.
Hi, I wonder if it is possible that pymlir takes .td files as input and parse them to Dialects? So that I don't have to hand write the python Dialects....
I am currently using pymlir to parse an mlir file with quite a bit of custom dialects used in the mlir file. A file ~2200 lines takes about 5 mins...
The generated MLIR does not have dialect specification. For example, **tests/test_builder.py** generates instructions without dialect (as shown below): **func -> should be -> func.func constant -> should be -> arith.constan**t...
 Is the function call here supposed to be like this 
```bash (.venv) vscode ➜ /workspaces/project (main) $ python3 -m project ./test/fibonacci/fibonacci.scf.mlir Traceback (most recent call last): File "", line 198, in _run_module_as_main File "", line 88, in _run_code File "/workspaces/project/project/__main__.py",...
Hi, I noticed that in the following example, the `#llvm.linkage` (pretty dialect attribute) is being parsed as a pretty dialect type (`!llvm.linkage`). I'm just curious— is this the expected behavior?...
Currently, LoadOperation and StoreOperation expect the instructions to just be "load" and "store", respectively. This is not possible in current MLIR - indeed, I don't know if it was ever...
