Skip to content
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

[GPU] Memory consumption optimizations backport to 24.3 #26076

Conversation

…tations (openvinotoolkit#25756)

### Details:
- This PR adds an ability to reuse single kernel between multiple
implementations, resulting in much lower memory pressure in some cases.
It can be enabled with `ov::intel_gpu::hint::enable_kernels_reuse(true)`
property
 
 ### Tickets:
 - [CVS-145296](https://jira.devtools.intel.com/browse/CVS-145296)
…rder kernels (openvinotoolkit#26025)

### Details:
- Currently, b_fs_yx_fsv16 blocked format is enabled for convolution
operations, but the lack of it for reorder and quantize operations
causes runtime static version kernel recompilation. This change enables
support for the b_fs_yx_fsv16 format to allow shape-agnostic kernel
selection.

### Tickets:
 - [CVS-145296](https://jira.devtools.intel.com/browse/CVS-145296)
@sshlyapn sshlyapn added the category: GPU OpenVINO GPU plugin label Aug 15, 2024
@sshlyapn sshlyapn added this to the 2024.3 milestone Aug 15, 2024
@sshlyapn sshlyapn requested review from a team as code owners August 15, 2024 06:07
@github-actions github-actions bot added category: inference OpenVINO Runtime library - Inference category: CPP API OpenVINO CPP API bindings labels Aug 15, 2024
@ilya-lavrenov
Copy link
Contributor

should be closed?

@p-durandin p-durandin closed this Aug 29, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
category: CPP API OpenVINO CPP API bindings category: GPU OpenVINO GPU plugin category: inference OpenVINO Runtime library - Inference
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants