ZLUDA v2 Released For Drop-In CUDA On Intel Graphics

One of many interesting and original open-source projects to be started in 2020 was ZLUDA, an open-spurce drop-in CUDA implementation for Intel graphics. ZLUDA - developed independent of Intel and NVIDIA - is built atop Intel's oneAPI Level Zero interface (hence the name, ZLUDA) and allows for unmodified CUDA applications to run on Intel UHD/Xe Graphics hardware with near-native performance. Well, that's the goal at least but with the initial ZLUDA release were a number of support limitations.

Out today is ZLUDA Version 2 that has been focused on ensuring it works well with the Geekbench CUDA test cases as an interesting stressor for CUDA on Intel graphics. Additionally, the Microsoft Windows support for ZLUDA has been improved while continuing to provide first-rate Linux support.

ZLUDA Version 2 implements more host-side functions, a number of bugs have been fixed, I/O performance improvements, support for creating traces of CUDA kernel executions, and other enhancements for this open-source CUDA implementation built on oneAPI Level Zero.

Linux and Windows binary downloads along with more information on the new ZLUDA release can be found via the project's GitHub. Geekbench CUDA tests have been one of the main programs focused on testing the ZLUDA capabilities while it will be interesting to give ZLUDA v2 a whirl to see what more can run atop Intel hardware for Skylake/Gen9 graphics and newer.