-
Producing macros with reserved names may yield problems. These two lines can create problems with hip-clang (and comgr).
- https://github.com/ROCmSoftwarePlatform/MIOpen/blob/14d51c85df366e54de7329…
-
I am currently porting a CUDA application that uses threadfence and threadfence_block to enforce ordering. According to the HIP documentation (http://rocm-documentation.readthedocs.io/en/latest/Progr…
-
Since the updated deb files pushed to the repo, I seem to have to start declaring "\_\_device__" on functions where before "it just worked"?
Is there some justification for this?
-
I tried to run hipify-clang on a simple program (attached for reference). Interac doesn't have any reference calls, and hipify-perl is able to convert it perfectly except that it uses hipLaunchKernel…
-
Does win10 bash support GPU devices?
-
Current `ALPAKA_FN_*` macros:
```C
#if BOOST_LANG_CUDA
#define ALPAKA_FN_ACC_CUDA_ONLY __device__
#define ALPAKA_FN_ACC_NO_CUDA __host__
…
-
@jjhelmus it seems you were able to build pytorch GPU without needing to have variants
https://anaconda.org/anaconda/pytorch/files?version=1.0.1
Is that true?
If so, what challenges do you se…
-
This is an idea I've been thinking about recently. It isn't necessarily something we want to do right away. I just want to record my thoughts and see what other people have to say about it.
Large…
-
### Bug description
An [archived repo of mine](https://github.com/mdeff/ntds_2016) requires python 3.5 (specified in `runtime.txt`) and numpy 1.11.1 (specified in `requirements.txt`). While it used…
mdeff updated
4 years ago
-
I am currently evaluating different approaches to / frameworks for GPGPU programming in the context of HPC - namely SYCL, ROCm (HC / C++AMP and HIP) and CUDA. In order to test SYCL (via Codeplay's Com…