-
Notifications
You must be signed in to change notification settings - Fork 17
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Added the new GPU VA context class to handle VA connections to GPU. Removed the hard coded HW decoder capability info. #415
Conversation
jeffqjiangNew
commented
Aug 29, 2024
- We now probe HW decoder capabilities through VA-API from the driver.
- We now probe HW decoder capabilities through VA-API from the driver.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@jeffqjiangNew : let us meet to go over these changes sometime
@jeffqjiangNew needs merge conflicts resolved to run CI |
… to handle HIP and VA initialization and VA attributes probe for both HW capability check and decoder initialization.
…oder.h and into a new file. Removed debug logs. Removed roc_decoder_caps.h.
…led to vaInitialize() and vaTerminte() should be paired.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Added few comments.
@lakshmi to test with multiple instances
Testing with higher number of threads = 256 on perf app passes with driving video:
Testing with higher number of threads = 512 on perf app passes with driving video:
Testing with higher number of threads = 1024 on perf app passes with driving video:
|
Testing with multiple instances = 30 on videodecode app passes with driving video
Testing with multiple instances =100 on videodecode app passes with driving video
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Multi-instance on multi-GPU passes too