-
CUDA 9 `__shfl_sync` function is missing. I can use the deprecated `__shfl` but it would be
be better to have the new function.
Test code:
```
__global__
static void shflTest(int lid){
…
-
### Checklist
- [X] I have searched for [similar issues](https://github.com/isl-org/Open3D/issues).
- [X] For Python issues, I have tested with the [latest development wheel](https://www.open3d.org/d…
-
## Issue description
I want to build libtorch for rocm from source and followed the instructions from here:[Building libtorch using CMake](https://github.com/pytorch/pytorch/blob/main/docs/libtorch…
-
**[IMP]**
`int64_t` rocBLAS APIs should be available in the same compilation unit. That means, `int`-based APIs can `live` in the same code simultaneously with `int64_t` ones. For instance:
```cpp…
-
### Problem Description
When running hipify-clang on a cu file that imports cub.cuh I get the following error. Happens with CUDA 11.7 and 11.8.
```
In file included from /usr/local/cuda-11.7/incl…
-
Having a kernel launch inside a macro like this
``` C++
#define CUDA_LAUNCH(cuda_call,wthr,thr, ...) \
cuda_call(__VA_ARGS__);
```
and using it with
``` C++
CUDA_LAUNCH((count_nn_cells)…
-
susie.sun@yz-amd1:~$ docker run -it rocm/deepspeed:rocm5.7_ubuntu20.04_py3.9_pytorch_2.0.1_DeepSpeed /bin/bash
root@c50e90963e1a:/var/lib/jenkins# deepspeed --num_gpus 1 deploy.py
[2023-12-14 01:52:…
-
I have finally managed to put AMD Instinct card to work with the latest ROCm on my IBM POWER9 server.
I have also been trying to build xmrig with opencl support, however it seems that even the "por…
-
when i try to create the environment,it said No module named 'torch.utils.hipify',how can i deal with it?
-
Right now, it seems to suppress the error and keep going. That's bad! If I fail to write out a file, the subsequent build will almost certainly fail.
cc @malfet @seemethere @jeffdaily @sunway513 @jit…