Intel Compute Runtime 24.52.32224.5 Brings More Battlemage Optimizations
The Intel Compute Runtime 24.52.32224.5 release was made earlier today as the newest update to this open-source Linux and Windows compute stack for Intel graphics hardware for providing OpenCL and oneAPI Level Zero implementations.
With this being the first new version in about one month, it brings various new features and optimizations -- including for the recently launched Battlemage / Xe2 discrete graphics hardware. This v24.52.32224.5 release has additional Battlemage performance optimizations like enabling direct submission on Linux and timestamp wait for queues.
The updated Intel open-source GPU compute stack also brings a new API for reading the Vertex Fetch (VF) engine utilization, refactoring so the code can compile with C++20, new prefetch capabilities to Level Zero, and experimental support for external semaphores.
For downloading the source code or pre-built Ubuntu Linux binaries of the updated Intel Compute Runtime, visit the Intel GitHub page.
With this being the first new version in about one month, it brings various new features and optimizations -- including for the recently launched Battlemage / Xe2 discrete graphics hardware. This v24.52.32224.5 release has additional Battlemage performance optimizations like enabling direct submission on Linux and timestamp wait for queues.
The updated Intel open-source GPU compute stack also brings a new API for reading the Vertex Fetch (VF) engine utilization, refactoring so the code can compile with C++20, new prefetch capabilities to Level Zero, and experimental support for external semaphores.
For downloading the source code or pre-built Ubuntu Linux binaries of the updated Intel Compute Runtime, visit the Intel GitHub page.
3 Comments