- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I’m trying to build a GPU zero copy inference environment.
Refer to official documents from https://docs.openvino.ai/2022.3/openvino_docs_install_guides_installing_openvino_from_archive_linux.html, and install relevant GPU inference plugins according to the GPU setup guide to execute OpenVINO.
First, download static resource decompression from the website (OpenVINO 2022.3 or 2023.0).
Second, set environment variables (source /opt/intel/openvino/setupvars.sh ).
Finally, use the apt-get command on Ubuntu 20.04 or download 22.35.24055 from the website to install the GPU plugin.
Write a zero copy demo to test, but the underlying function interface reported an error (create_context). From the ldd command, it can be seen that all relevant dynamic library resources exist. (look at ldd_demo.png and gdb_demo.png)
The same situation occurs with the official resources of oneAPI. (MFXCreateSession error) (look at ldd_githubDemo.png and githubDemo.png)
GitHub: https://github.com/OpenVINO-dev-contest/decode-infer-on-GPU
Looking forward to your reply, thanks.
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
Could you share:
- What GPU hardware are you using (eg: Intel UHD, Intel Xeon, etc)
- Relevant model files
- Relevant custom code
Cordially,
Iffa
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
Thank you for your question. If you need any additional information from Intel, please submit a new question as Intel is no longer monitoring this thread.
Cordially,
Iffa
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page