1
[Feature] Rusticl Support for Machine Learning Hardware Acceleration
Source: immich-app/immich#30008 · opened by @hopechurch1
I have searched the existing feature requests, both open and closed, to make sure this is not a duplicate request.
• Yes
The feature
The latest versions of ROCm don't support older AMD GPUs (i.e. Polaris, Vega, older RDNA) which are still supported by Mesa Rusticl. Kernel & OS compatibility with ROCm is also somewhat flaky in comparison to Rusticl.
Supporting hardware accelerated machine learning on these GPUs would allow faster and more efficient processing of large image libraries.
Platform
• Server
• Web
• Mobile
• Yes
The feature
The latest versions of ROCm don't support older AMD GPUs (i.e. Polaris, Vega, older RDNA) which are still supported by Mesa Rusticl. Kernel & OS compatibility with ROCm is also somewhat flaky in comparison to Rusticl.
Supporting hardware accelerated machine learning on these GPUs would allow faster and more efficient processing of large image libraries.
Platform
• Server
• Web
• Mobile
No pledges yet. Be the first to back this.
Comments
Similar requests
[Feature] Support Hailo for machine learning hardware acceleration
34 votes · 0 comments
[Feature] DirectML execution provider for Machine Learning (AMD GPU Support on Windows/WSL2)
8 votes · 0 comments
[Feature] Separate Image Preview for Machine Learning
1 vote · 0 comments
Improve machine learning resilience in low VRAM scenarios (CUDA)
1 vote · 0 comments
[Feature] Support for arm64 for CUDA (immich-machine-learning)
41 votes · 0 comments
No comments yet.