Packages included in the release
- intel-gmmlib (https://github.com/intel/gmmlib)
- intel-opencl-icd, intel-level-zero-gpu (https://github.com/intel/compute-runtime)
Components revisions included in the release
Additional components revisions used in build
- Used for building runtime
- intel/[email protected] (Compatible with va_api_major_version = 1)
- oneapi-src/[email protected] (Supports oneAPI Level Zero Specification v1.12.15)
- intel/[email protected]
- intel/[email protected]
Installation procedure on Ubuntu 22.04/24.04
Please be aware, since next release we are going to upgrade our packages from Ubuntu 22.04 to Ubuntu 24.04.
If you are using legacy platform, follow latest 24.35 release procedure instead the one below. More details in known issues section.
- Create temporary directory
Example:
mkdir neo
- Download all *.deb packages
Example:
cd neo
wget https://github.com/intel/intel-graphics-compiler/releases/download/v2.10.8/intel-igc-core-2_2.10.8+18926_amd64.deb
wget https://github.com/intel/intel-graphics-compiler/releases/download/v2.10.8/intel-igc-opencl-2_2.10.8+18926_amd64.deb
wget https://github.com/intel/compute-runtime/releases/download/25.13.33276.16/intel-level-zero-gpu-dbgsym_1.6.33276.16_amd64.ddeb
wget https://github.com/intel/compute-runtime/releases/download/25.13.33276.16/intel-level-zero-gpu_1.6.33276.16_amd64.deb
wget https://github.com/intel/compute-runtime/releases/download/25.13.33276.16/intel-opencl-icd-dbgsym_25.13.33276.16_amd64.ddeb
wget https://github.com/intel/compute-runtime/releases/download/25.13.33276.16/intel-opencl-icd_25.13.33276.16_amd64.deb
wget https://github.com/intel/compute-runtime/releases/download/25.13.33276.16/libigdgmm12_22.7.0_amd64.deb
- Verify sha256 sums for packages
Example:
wget https://github.com/intel/compute-runtime/releases/download/25.13.33276.16/ww13.sum
sha256sum -c ww13.sum
- Install all packages as root
Example for Ubuntu 22.04:
sudo dpkg -i *.deb
In case using Ubuntu 24.04 our packages may conflict with Ubuntu 24.04 packages: intel-ocloc-dev intel-ocloc libze-intel-gpu1. To resolve this, you may remove those packages, as in given example.
Example for Ubuntu 24.04:
sudo dpkg -r intel-ocloc-dev intel-ocloc libze-intel-gpu1
sudo dpkg -i *.deb
In case of installation problems, please install required dependencies, for example:
apt install ocl-icd-libopencl1
sha256 sums for packages
d0ffa3219623486901b3dd752c4d74afe79320471713b7772777c7319e58b5f7 intel-level-zero-gpu-dbgsym_1.6.33276.16_amd64.ddeb
b7dbea8eac25cac46bca1768f356626f22f40d0ac9ccd80e0c13402ddc573262 intel-level-zero-gpu_1.6.33276.16_amd64.deb
1a482fa033963291e9ed4596063ff2d156aedebe4379ff33d9a28baa5dfe9cc3 intel-opencl-icd-dbgsym_25.13.33276.16_amd64.ddeb
db5245e061a2eb043af655f298b045a4ead5973b8e5d8f385a1d2887deb309b0 intel-opencl-icd_25.13.33276.16_amd64.deb
1c20bcbfbc3cc293e3a2ef7f8876ae771e0b9458f95a13bc80d2d69cf0bbef23 libigdgmm12_22.7.0_amd64.deb
For IGC related sha256 packages sums see IGC release
Quality expectations
| Platform | Quality | OpenCL | Level Zero | WSL |
|---|---|---|---|---|
| DG1 | Production | 3.0 | 1.6 | Yes |
| Alchemist | Production | 3.0 | 1.6 | Yes |
| Battlemage | Production | 3.0 | 1.6 | Yes |
| Tiger Lake | Production | 3.0 | 1.6 | Yes |
| Rocket Lake | Production | 3.0 | 1.6 | Yes |
| Alder Lake | Production | 3.0 | 1.6 | Yes |
| Meteor Lake | Production | 3.0 | 1.6 | Yes |
| Raptor Lake | Production | 3.0 | 1.6 | Yes |
| Lunar Lake | Production | 3.0 | 1.6 | Yes |
| Arrow Lake | Production | 3.0 | 1.6 | Yes |
| Panther Lake | Pre-Release | 3.0 | 1.6 | Yes |
For Lunar Lake and Battlemage platforms the best experience is achieved via using Ubuntu 24.10 with intel-graphics PPA. See instructions
For legacy platforms not included in this table, see more information.
All platforms were validated on Ubuntu 24.04 LTS with stock kernel 6.8.0-48-generic, unless noted otherwise.
- Lunar Lake and Battlemage was tested using Ubuntu 24.10 with 6.11.0-1007-intel kernel from intel-graphics PPA.
WSL support was tested for all platforms with Windows host driver 101.6732, unless noted otherwise.
Quality levels
- Experimental - no quality expectations
- Early support - platform may not be available on the market yet
- Pre-Release - suitable for integration and testing, gathering customer feedback
- Beta - suitable for integration and broad testing
- Production - Beta + meets API-specific conformance requirements; suitable for production
- Maintenance - Reduced test frequency compared to Production, no longer recommended for new deployments. Reactive support for customers.
Additional information
- packages were built with custom flags NEO_ENABLE_I915_PRELIM_DETECTION=1