How to quickly switch between using multiple ICudaEngines when GPU memory is limited?
Open
Nobody has claimed this yet.
triaged
- Dominant language
- C++
- Stars
- 13.4k
- Forks
- 2.4k
- Avg merge
- 5d 3h
- Merged PRs (30d)
- 2
Description
Description
My current approach is to use a class TRT to encapsulate nvifer1:: IExecutionContext and other objects together, and then deconstruct the TRT. When needed again, I use deserialization to build the engine, and then build nvinfer1::IExecutionContext。Is there a better and faster way？
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
The issue names no repository files, tests, or entry points. Start by reviewing the described TRT wrapper, ICudaEngine deserialization, and IExecutionContext lifecycle; done means identifying and documenting a faster supported approach for switching engines under GPU-memory limits.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- machine-learning, performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100