Add ability to update kernel parameters of a captured graph
Open
Nobody has claimed this yet.
feature request
runtime
- Dominant language
- Python
- Stars
- 7.1k
- Forks
- 624
- Avg merge
- 3d 17h
- Merged PRs (30d)
- 5
Description
See cudaGraphExecKernelNodeSetParams: https://developer.nvidia.com/blog/constructing-cuda-graphs-with-dynamic-parameters/
Reuse command object API so that set_param() can be used to update the kernel parameters of captured graphs?
- During graph capture, a
Launchobject would be returned fromwp.launch()that can be used to update the kernel parameters after the graph is captured.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with the linked CUDA documentation for cudaGraphExecKernelNodeSetParams, then inspect the existing command object API around wp.launch() and set_param(). Determine how a Launch object from graph capture should expose kernel-parameter updates, and define completion around updating captured graph parameters through that API.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- api, backend
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100