Show Your Support: This site is primarily supported by advertisements. Ads are what have allowed this site to be maintained on a daily basis for the past 18+ years. We do our best to ensure only clean, relevant ads are shown, when any nasty ads are detected, we work to remove them ASAP. If you would like to view the site without ads while still supporting our work, please consider our ad-free Phoronix Premium.
Intel Compute Runtime Preparing For The Upcoming oneAPI Level Zero 1.0
At the end of last year Intel published the oneAPI Level Zero specification as a low-level API for direct-to-metal interfaces for offload accelerators like FPGAs and GPUs. In the months since they have continued advancing the Level Zero interface and implementation within the Intel software stack (along with the other oneAPI components at large) while it's looking like Level Zero v1.0 is around the corner.
Currently the latest advertised release is v0.91.21. Meanwhile the documentation points to a 1.0.4 version.
But what caught my attention today was with the Intel Compute Runtime 20.31.17495 weekly update. Aside from adding new Gen11 PCI IDs and updating the Intel Graphics Compiler (IGC) and memory management (GMMLIB), there is also Level Zero 1.0 work that isn't mentioned as part of the release notes.
Officially the Intel Compute Runtime stack for Linux exposes Level Zero 0.8, but when looking at the Git activity for this release over the past week there is indeed changes across multiple commits working on "level-zero v1.0."
Great to see that it looks like Level Zero 1.0 will likely be officially announced soon. It wouldn't surprise me at all if it happens to be timed with the upcoming rumored Tiger Lake hardware announcements. The Intel oneAPI software ecosystem has been advancing very well over the past year and is very exciting their numerous components not only for Level Zero and compute but over to OpenVKL, OIDn, OSPray, and other components. And it's all open-source!