# OpenCL example

**URL:** https://discourse.llvm.org/t/opencl-example/74187
**Category:** MLIR
**Created:** [October 17, 2023, 2:35pm UTC](https://discourse.llvm.org/t/opencl-example/74187 "2023-10-17T14:35:14Z")
**Posts on this page:** 5
**Page:** 1

<div class="post-metadata">

### Author: ![jungpark](https://sea1.discourse-cdn.com/flex021/user_avatar/discourse.llvm.org/jungpark/32/1495_2.png) [@jungpark](https://discourse.llvm.org/u/jungpark)
#### Post date: [October 17, 2023, 2:35pm UTC](https://discourse.llvm.org/t/opencl-example/74187/1 "2023-10-17T14:35:14Z")

</div>

Hello everyone,

I’d like to introduce a simple e2e MLIR example which runs on the GPU via OpenCL. (i.e., for non-nvidia GPUs)  
There are some existing solutions that use OpenCL runtime to run a GPU kernel generated by MLIR, but not many are available in public (yeah, I’ve just googled again ‘mlir opencl’ and got nothing new) . So, I’ve made an example to easily try.  
This doesn’t add any implementation to the MLIR core, instead MLIR compiles/passes opencl kernel binary and run it on python. It’s not a perfectly standard way and there could be more and better options.

> **[GitHub - jungpark-mlir/mlirPyoclExec: Enabling OpenCL in MLIR via Python](https://github.com/jungpark-mlir/mlirPyoclExec/tree/main)**
>
> main

There’s only one test available under `examples/` folder now, it simply adds up two 1024xf32 memrefs, using linalg.elemwise\_binary.  
It lowers and parallel maps to the GPU threads using upstream passes. And lowers down to the llvm dialect via rocdl dialect so it can be used by the AMD rocm OpenCL stack. (It shouldn’t be hard to make a spirv pipeline for the Intel GPU)

I’ve encountered an issue using `gpu-module-to-binary` pass in python[1], used `mlir-opt` host tool for the last step and had to manually convert the stringfied GPU binary in python. Now the GPU executable is obtained, all done from MLIR in this example.

Finally it creates an OpenCL kernel from the binary and runs it on the GPU with the Numpy in/output data using pyopencl. (inputs are random)

```auto
(mlirdev) $ python3 ./memrefAdd.ocl.bin.rocm.py 
A : [0.45386615 0.15114269 0.76626986 ... 0.8001651 0.37151784 0.60125226]
B : [0.69195175 0.74433124 0.36322945 ... 0.03635408 0.842453 0.9205129]
Validating A + B ...
Numpy : [1.1458179 0.89547396 1.1294993 ... 0.8365192 1.2139709 1.5217652]
GPU : [1.1458179 0.89547396 1.1294993 ... 0.8365192 1.2139709 1.5217652]
Pass : True

```

Hope this helps and any question is welcome here.

* * *

[1] `gpu-module-to-binary` pass fails via python binding while it runs fine with mlir-opt

```auto
     pm.run(module.operation)
mlir._mlir_libs._site_initialize.<locals>.MLIRError: Failure while executing pass pipeline:
error: unknown: cannot be converted to LLVM IR: missing `LLVMTranslationDialectInterface` registration for dialect for op: gpu.module

```

---

<div class="post-metadata">

### Author: ![mehdi\_amini](https://sea1.discourse-cdn.com/flex021/user_avatar/discourse.llvm.org/mehdi_amini/32/25595_2.png) [@mehdi\_amini](https://discourse.llvm.org/u/mehdi_amini)
#### Post date: [October 17, 2023, 9:09pm UTC](https://discourse.llvm.org/t/opencl-example/74187/2 "2023-10-17T21:09:15Z")

</div>

> [@jungpark](#):
>
> ```auto
> pm.run(module.operation)
> mlir._mlir_libs._site_initialize.<locals>.MLIRError: Failure while executing pass pipeline:
> error: unknown: cannot be converted to LLVM IR: missing `LLVMTranslationDialectInterface` registration for dialect for op: gpu.module
> 
> ```

The error is hinting at the issue “missing `LLVMTranslationDialectInterface`”, the `mlir-opt` tool is registering this translation interface but your python environment does not. This is done by calling `mlir::registerGPUDialectTranslation`.

---

<div class="post-metadata">

### Author: ![jungpark](https://sea1.discourse-cdn.com/flex021/user_avatar/discourse.llvm.org/jungpark/32/1495_2.png) [@jungpark](https://discourse.llvm.org/u/jungpark)
#### Post date: [October 23, 2023, 1:23pm UTC](https://discourse.llvm.org/t/opencl-example/74187/3 "2023-10-23T13:23:51Z")

</div>

It doesn’t seem there’s an existing code path to `mlir::registerGPUDialectTranslation` from python binding at the moment.

Some random findings,

1. [Registereverything](https://github.com/llvm/llvm-project/blob/324d1bb35aefec737ef381d98211f7771c4b0ab5/mlir/lib/Bindings/Python/RegisterEverything.cpp#L15) for the python binding looks different from what we can do in C-API `mlir/lib/CAPI/RegisterEverything/RegisterEverything.cpp`. It lacks of calling `mlirRegisterAllLLVMTranslations` but not sure if that’s the intended behaviour. Also calls to `mlirRegisterConversionPasses()` and `mlirRegisterTransformsPasses()` are redundant since underneath functions are already called in `mlirRegisterAllPasses()`.

2. I felt bit hacky while unsure where’s the best place for the function, tried to directly call C-API function from python as below and it works.

```auto
import os
import ctypes
from ctypes import *
from mlir._mlir_libs import get_lib_dirs
capi_so = os.path.join(get_lib_dirs()[0], "libMLIRPythonCAPI.so")
capi_functions = CDLL(capi_so)

ctx = Context()
ctx_capsule = ctx._CAPIPtr

MLIR_PYTHON_CAPSULE_CONTEXT = c_char_p("mlir.ir.Context._CAPIPtr".encode())
ctypes.pythonapi.PyCapsule_GetPointer.restype = ctypes.c_void_p
ctypes.pythonapi.PyCapsule_GetPointer.argtypes = [ctypes.py_object, ctypes.c_char_p]
ctx_ptr = ctypes.pythonapi.PyCapsule_GetPointer(ctx_capsule, MLIR_PYTHON_CAPSULE_CONTEXT)
capi_functions.mlirRegisterAllLLVMTranslations.argtypes = [ctypes.c_void_p]
capi_functions.mlirRegisterAllLLVMTranslations(ctx_ptr)

```

1. I can’t clearly understand the usage of the `_CAPIPtr` capsule. All tests only check if they can recreate the same object from the capsule, I suppose that’s not all we can do with it. Is above case included in the expected use case?

---

<div class="post-metadata">

### Author: ![mehdi\_amini](https://sea1.discourse-cdn.com/flex021/user_avatar/discourse.llvm.org/mehdi_amini/32/25595_2.png) [@mehdi\_amini](https://discourse.llvm.org/u/mehdi_amini)
#### Post date: [October 24, 2023, 5:30am UTC](https://discourse.llvm.org/t/opencl-example/74187/4 "2023-10-24T05:30:04Z")

</div>

It’s likely that “register everything” should also register the translations, feel free to send a PR!

---

<div class="post-metadata">

### Author: ![jungpark](https://sea1.discourse-cdn.com/flex021/user_avatar/discourse.llvm.org/jungpark/32/1495_2.png) [@jungpark](https://discourse.llvm.org/u/jungpark)
#### Post date: [October 27, 2023, 1:50pm UTC](https://discourse.llvm.org/t/opencl-example/74187/5 "2023-10-27T13:50:36Z")

</div>

Hope it’s in the correct place.  
[Register all translations in the RegisterEverything for python by jungpark-mlir · Pull Request #70428 · llvm/llvm-project (github.com)](https://github.com/llvm/llvm-project/pull/70428)