intel-deep-learning-essentials

Intel® Deep Learning Essentials + oneDNN, packaged for Arch Linux.

This is a dependency package for llama.cpp-sycl. It installs only the oneAPI components actually needed to run llama.cpp with SYCL acceleration, the compiler, MKL, oneDNN, CCL, and the DPC++ library, instead of the full bloated and/or outdated intel-oneapi-toolkit, intel-oneapi-basekit-2025 or intel-oneapi-hpckit packages.

Installs to /opt/intel/oneapi/ like in the official docs.

If the install fails

If you encounter issues while installing this package, 99% of the time it's one of the following options:

• If you previously or currently have the intel-oneapi-toolkit, intel-oneapi-basekit-2025 or intel-oneapi-hpckit packages installed, please uninstall them and verify that everything in /opt/intel is deleted. Most likely you don't even need everything that's included in the full Intel® oneAPI Toolkit for developers, that's why the Intel® Deep Learning Essentials for users exists. I suggest replacing the three packages with this one unless you are one of the few people that need the full kit.
• If you are updating this package and it fails, you have to delete /home/yourusername/intel and /home/yourusername/.intel and retry the update. Please verify just in case, but there usually is nothing important in there anyway.

Why this exists

The AUR/extra has the full intel-oneapi-toolkit, intel-oneapi-basekit-2025 or intel-oneapi-hpckit packages with all the other analysis-, debug tools and add-ons which is gigabytes of stuff you simply don't need. The full Intel® oneAPI Toolkit is intended for developers, if all you want is SYCL acceleration for llama.cpp, you only need a small subset of it. That's why the Intel® Deep Learning Essentials exist. This package ships exactly that and nothing else for close to 10 GB less than the full kit. It also bundles oneDNN, which is available on the arch repos already, but it is essential to download the matching package to the Intel® Deep Learning Essentials to avoid issues in the /opt/intel/oneapi/ folder, which is why I decided to bundle it here instead of adding it as a dependency. OneDNN used to be part of the Intel® Deep Learning Essentials up until 2025.3.3, no idea why they even decided to ship it seperately with 2026.0.0.

Install

Via AUR helper

yay -S intel-deep-learning-essentials
# or
paru -S intel-deep-learning-essentials

Manual

git clone https://github.com/cantosun99/intel-deep-learning-essentials.git
cd intel-deep-learning-essentials
makepkg -si

The installers are downloaded directly from Intel and are large. This is expected.

Notes

• The package must run outside of fakeroot so the Intel installer respects --install-dir. This is intentional and handled in the PKGBUILD.
• A linker config is written to /etc/ld.so.conf.d/ so the compiler libs are found system-wide without manual environment setup.
• Installer logs are removed from the final package to avoid $srcdir/$pkgdir path leakage.

License

PKGBUILD and packaging: MIT. Intel oneAPI components are subject to Intel's license terms.