Questions about Intel Packages
Hi, I'm working on a youtube video about which Intel packages to install on Ubuntu, Debian, and potentially openSUSE to be able to use an Intel Arc/Xe Core APU for compute functions. I had some questions about the structure and choices of these packages, since I noticed some differences between how things are done for Debian/Ubuntu, and other things. If this isn't the best place to ask these questions, please direct me to the right mailing list or forum. First Question: On Ubuntu, Libmfx-gen1.2 is the only package needed for oneVPL-intel-gpu, while on openSUSE Tumbleweed, you need libmfx-gen and libmfx-gen_2 for seemingly the same functionality? Libmfx-gen is 6.6KiB while libmfx-gen1_2 is 8.4MiB, could someone explain why there's a size discrepancy, and what the function of the two packages are? Second Question: I noticed that libvp2 is being built based on oneVPL 2.10.1-2.2, while 2.10.1 was released in December 2023. 2.10.1 was released in December 2023, and that's the only 2.10.1 I see available on github. Are versions like 2.10.1-2.2 created by the openSUSE maintainers with adjustments for openSUSE or... where do they come from? Sub question The latest version is 2.11.0, and I'm wondering if there's an issue as far as it not being the version available for openSUSE Tumbleweed users? Are the intel packages not a priority, especially since there's uncertainty about Intel's gpu future? 3rd question I've tried running the OpenVINO implementation of Stable Diffusion WebUI, and all of the packages needed seem to be available on the openSUSE repo, but script doesn't complete. Would you know if this is because there are at least some packages that are behind the latest of what Intel has made available version wise, or is it an issue with the parameters being passed during compilation by the openSUSE maintainers. I'd really appreciate it if I could get some answers so that I can report a straight story to my audience because I'd really like for Tumbleweed (at least) to be a platform (I'm a Tumbleweed user) that people could utilize Intel Arc/Xe cores on for compute/productivity/AI tasks, instead of having to dual boot into Ubuntu. I hope it doesn't seem that I'm just trying to stick my nose into your business for my own gain; I'd genuinely like to help more people be able to use openSUSE (Tumbleweed) for as many things as possible. Thanks!
On 28.06.2024 10:43, compellingbytes@protonmail.com wrote:
Second Question: I noticed that libvp2 is being built based on oneVPL 2.10.1-2.2, while 2.10.1 was released in December 2023. 2.10.1 was released in December 2023, and that's the only 2.10.1 I see available on github. Are versions like 2.10.1-2.2 created by the openSUSE maintainers with adjustments for openSUSE or... where do they come from?
Release numbers are auto-generated by OBS (the first number is commit count and the second number is rebuild count).
On Friday 2024-06-28 09:43, compellingbytes@protonmail.com wrote:
First Question: On Ubuntu, Libmfx-gen1.2 is the only package needed for oneVPL-intel-gpu, while on openSUSE Tumbleweed, you need libmfx-gen and libmfx-gen_2 for seemingly the same functionality? Libmfx-gen is 6.6KiB while libmfx-gen1_2 is 8.4MiB, could someone explain why there's a size discrepancy, and what the function of the two packages are?
have you looked at the package descriptions? package contents/filelist?
Sub question The latest version is 2.11.0, and I'm wondering if there's an issue as far as it not being the version available for openSUSE Tumbleweed users? Are the intel packages not a priority, especially since there's uncertainty about Intel's gpu future?
it's a community, not a job, people do whatever and whenever they feel like it.
Description for libmfx-gen: libmfx-gen - Intel oneVPL GPU Runtime Intel oneVPL GPU Runtime is a Runtime implementation of oneVPL API for Intel Gen GPUs: Runtime provides access to hardware-accelerated video decode, encode and filtering. Description for libmnfx-gen1_2: libmfx-gen1_2 - Intel oneVPL GPU Runtime Intel oneVPL GPU Runtime is a Runtime implementation of oneVPL API for Intel Gen GPUs: Runtime provides access to hardware-accelerated video decode, encode and filtering.
To ensure yourself high grades, you can turn to https://papercheap.co.uk/assignment-writing-service which always provides an individual approach to each order. You can be sure that your assignment will be completed with the utmost care and accuracy. This allows students to save time and effort, which they can devote to other important aspects of their academic life. Pleasant price, fast execution and many other benefits will be available to you.
participants (4)
-
Andrei Borzenkov
-
compellingbytes@protonmail.com
-
Jan Engelhardt
-
marlinpavlin@gmail.com