Package Details: openvino-intel-gpu-plugin 2026.2.0-3

Git Clone URL: https://aur.archlinux.org/openvino.git (read-only, click to copy)
Package Base: openvino
Description: Intel GPU plugin for OpenVINO
Upstream URL: https://docs.openvino.ai/
Licenses: Apache-2.0, LicenseRef-custom
Submitter: dbermond
Maintainer: dbermond
Last Packager: dbermond
Votes: 17
Popularity: 0.101952
First Submitted: 2020-12-23 22:29 (UTC)
Last Updated: 2026-06-04 21:04 (UTC)

Required by (2)

Sources (27)

Latest Comments

1 2 3 4 5 6 .. 12 Next › Last »

leenux commented on 2026-06-07 19:50 (UTC)

Building now. Tried several times before. Probably missing dependencies: gst-plugin-onnx python-pip python-uv python-virtualenv

matthias.lisin commented on 2026-06-05 16:36 (UTC)

@dbermond You're right. Thank you for clarifying.

dbermond commented on 2026-06-05 16:22 (UTC) (edited on 2026-06-05 16:25 (UTC) by dbermond)

@matthias.lisin The number of build jobs in makepkg is a user configuration, and not a package setting. Enforcing this in the PKGBUILD is not the way to go, unless when explicitly needed (like in the cases when a single build job is required). This should be configured by setting the 'MAKEFLAGS' variable in your makepkg.conf.

matthias.lisin commented on 2026-06-05 15:52 (UTC)

@dbermond Thanks for maintaining this monster of a package! Can you look into adding --parallel $(nproc) to the cmake --build command? In my case ($(nproc) -eq 24) it takes significantly less time to build. Like about an hour less.

dbermond commented on 2026-06-04 21:08 (UTC)

@leenux I could reproduce your issue. It happens when having the onnx package installed. It can be circumvented by building the package in a clean chroot, but, for convenience, I added a patch for ignoring system onnx. It should be building fine now when having system onnx installed.

leenux commented on 2026-06-02 17:41 (UTC)

still not building: CMake Error at /usr/lib/cmake/ONNX/ONNXTargets.cmake:70 (set_target_properties): The link interface of target "ONNX::onnx" contains: 

absl::absl_check

selebray1998 commented on 2026-05-23 21:45 (UTC) (edited on 2026-05-23 21:48 (UTC) by selebray1998)

Well, I'm now (trying) to maintain audacity-openvino. I'm new to this so I'm certain that I messed a few things up. But I am trying day by day to get it to build, so bear with me. Shouldn't be much longer now. By the way, is there any plans to implement GPU acceleration support for these packages? You can look at my packages for conditional logic that checks to see if there's a GPU/compiler and builds for the correct one and if it doesn't detect one it falls back to CPU.

alzpmq commented on 2026-05-20 13:24 (UTC)

The PKGBUILD will generate /usr/lib/libopenvino_jax_frontend.so.2026.1.0.dbg,which make ldconfig segmentfault

newfstatat(AT_FDCWD, "/usr/lib/libopenvino_jax_frontend.so.2026.1.0", {st_mode=S_IFREG|0755, st_size=532296, ...}, AT_SYMLINK_NOFOLLOW) = 0
newfstatat(AT_FDCWD, "/usr/lib/libopenvino_jax_frontend.so.2026.1.0.dbg", {st_mode=S_IFREG|0644, st_size=3648, ...}, AT_SYMLINK_NOFOLLOW) = 0
openat(AT_FDCWD, "/usr/lib/libopenvino_jax_frontend.so.2026.1.0.dbg", O_RDONLY) = 4
fstat(4, {st_mode=S_IFREG|0644, st_size=3648, ...}) = 0
mmap(NULL, 3648, PROT_READ, MAP_SHARED, 4, 0) = 0x7f17af548000
--- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x7f1851c75a94} ---
+++ killed by SIGSEGV (core dumped) +++
[1]    199467 segmentation fault (core dumped)  strace -f ldconfig -r / 2>&1 | 
       199468 done                              tail -n 50

leenux commented on 2026-05-01 20:13 (UTC)

and still not building here.

leenux commented on 2026-04-10 09:01 (UTC) (edited on 2026-04-15 08:46 (UTC) by leenux)

And I forgot: audacity-openvino and audacity-openvino-bin won´t build either anymore because of libtorch-cxx11abi-cpu. So, without audacity-openvino, this all doesn´t make any sense, neither. We call it asymmetric warfare. Every community gets indoctrinated.

1 2 3 4 5 6 .. 12 Next › Last »