Leo Fang
We have examples like this today ```python kernel = module.get_kernel("vectorAdd") ``` which does not tell us what args are expected on the device side, and so when `launch(s, config, kernel,...
After #1260 I expect that this commit should clean up all contents instead of making minor changes: https://github.com/NVIDIA/cuda-python/commit/ff737c4c181d0c6fc97401075143a6d3c8b6b137
After #1242 is merged, we have a nice test matrix for different Windows configurations. However, currently they lock at the same driver version `latest`: https://github.com/NVIDIA/cuda-python/blob/c4079ddfb6b6dcdd7c9773b727530f122021fa9e/ci/test-matrix.json#L37-L48 But looking at this more...
Right now, we hard-code the backport branch name here: https://github.com/NVIDIA/cuda-python/blob/main/.github/BACKPORT_BRANCH We should relax it and allow: - triggering the backport workflow manually - providing a branch name to overwrite the...
Today before triggering the release workflow we need to create a git tag locally and push it to the remote repo, because 1. GitHub for whatever reason does not offer...
Tentative design considerations/constraints: 1. Must not run `cuInit(0)` or initialize CUDA if all users need is querying via NVML 2. Cover `nvmlDeviceGetCpuAffinity`, `nvmlDeviceGetMemoryInfo`, `nvmlDeviceGetCount`, ...
### Is this a duplicate? - [x] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cuda-python/issues) for this request and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.md) ### Area...