-
The error occurs when I am trying to install by
``
python -m pip install oneccl_bind_pt -f https://developer.intel.com/ipex-whl-stable-xpu
``
with Pytorch 1.13.0+cu117, Python 3.11.3, Ubuntu 22.0…
-
[https://github.com/merthidayetoglu/CommBench/blob/master/commbench.h](url)
Line 40 to 46 what is the point of commenting out the #define CAP_NCCL and others.
I have to manually uncomment it out…
-
## 🚀 Feature
The idea is to add "oneCCL" backend by extending our [torch native dist model](https://github.com/pytorch/ignite/blob/395743d1f782ca84a7fb1a3c71afe2fa76b882a1/ignite/distributed/comp_m…
-
**Describe the bug**
The builds on conda-forge have been failing since `deepspeed=0.14.1` for CUDA 11.8 and 12.0 with an error like `fatal error: oneapi/ccl.hpp: No such file or directory`. Origina…
-
Hi! I've installed `oneccl_bindings_for_pt==1.13.100+gpu` from https://developer.intel.com/ipex-whl-stable-xpu, but after installing I get a "libccl.so.1 not found" error:
```python
$ python …
-
It would be nice to be able to include oneCCL inside another build directly.
```
add_subdirectory(oneCCL)
```
This entails that all CMake configuration arguments should be prefixed by somethin…
-
**Environment:**
1. Framework: TensorFlow
2. Framework version: 2.6 and 2.8
3. Horovod version: 0.22.1 and 0.24.2
4. MPI version: OneCCL
5. CUDA version: N/A
6. NCCL version: N/A
7. Python vers…
-
**Environment:**
1. Framework: PyTorch
2. Framework version:1.6.0
3. Horovod version:0.21.3
4.oneCCL version:Intel(R) MPI Library for Linux* OS, Version 2021.2
5. CUDA version:10.1
6. Python ver…
-
Define processes for decision making within the oneDPL project. This should include a RFC process for new design and feature proposals
(see oneDNN for example - PR tagged RFC that links to the propos…
-
Define processes for decision making within the oneDAL project. This should include a RFC process for new design and feature proposals
(see oneDNN for example - PR tagged RFC that links to the propos…