Technology1 min read

NVIDIA CUDA 13.5 Brings NVIDIA Vera Rubin GA Support

By · Published by Everything Blog

In short

The software platform upgrade introduces comprehensive scheduling improvements through locality domain options and Multi-Process Service v3. Developers can now utilize weighted timeslicing, dynamic resource lending, and cgroup-based routing to balance shared GPU workloads efficiently. In addition to compiler and library updates, the release integrates batched 1D memory copies into reusable graph workflows without altering existing call syntax.

Key points

  • The update delivers general availability support for the NVIDIA Vera Rubin architecture i…: The update delivers general availability support for the NVIDIA Vera Rubin architecture in production environments.
  • Enhanced Multi-Process Service features optimize resource allocation and memory accountin…: Enhanced Multi-Process Service features optimize resource allocation and memory accounting in shared enterprise systems.
  • Batched 1D memory copies are now integrated directly into reusable CUDA Graphs workflows.: Batched 1D memory copies are now integrated directly into reusable CUDA Graphs workflows.

Sources checked by our analysis pipeline: 1

As artificial intelligence and accelerated computing continue to push the boundaries of what’s possible, the latest NVIDIA CUDA release arrives to unlock unprecedented performance and scale.

NVIDIA CUDA 13.5 delivers general availability (GA) across major new platforms. NVIDIA Vera Rubin support moves from early developer access to general availability (GA), ready for enterprise production. Windows on Arm support, GA since CUDA 13.4, now extends to NVIDIA DGX Station for Windows.

CUDA 13.5 also introduces more customization of how GPU work is scheduled via locality domains and NVIDIA Multi-Process Service (MPS). And CUDA 13.5 provides innovations across libraries, compilers, NVIDIA CUDA Tile, and developer tools.

This post will showcase the highlights of CUDA 13.5.

CUDA 13.5 enhancements

Some notable enhancements that come with CUDA 13.5 include:

Locality domain enhancements

CUDA 13.5 gives applications a choice between uniform streaming multiprocessor (SM) partitions across locality domains or partitions that include all available SMs. By default, localized green-context partitions use the same guaranteed SM count per domain, leaving supplemental SMs outside those partitions.

The new cudaDevSmResourceGroupLocalityDomainSupplementalSms

flag, combined with cudaDevSmResourceGroupLocalityDomainId

, includes these additional…

Technology