According to NVIDIA’s internal release calendar (viewed May 12):
The phased rollout is intentional. NVIDIA expects early bugs in the BME scheduler and UVM 2.5 prefetcher. They are letting AI labs and HPC centers test first before pushing to gamers. cuda driver release news exclusive
Two weeks ago, NVIDIA quietly pushed a new Production Branch driver to its developer portal without a typical blog post fanfare. Our analysis of the release notes (or lack thereof) reveals a build that is less about game-ready optimizations and entirely focused on two things: AI inference latency and virtualized memory paging. The phased rollout is intentional
Here is what the changelog doesn’t tell you: Two weeks ago, NVIDIA quietly pushed a new
# Use the developer beta runfile (leaked)
chmod +x cuda_570.85.05_linux.run
sudo ./cuda_570.85.05_linux.run --toolkit --samples --no-opengl-libs --no-man-page
Pro tip: Append --override if installing on unsupported kernel.
Buried in the R570 driver package is a new header file: cudaDriverExtension.h. It exposes three new functions that have never been publicly documented: