NVIDIA CUDA 13.4 released with memory-enforced GPU sharing feature
The story so far
NVIDIA has released CUDA 13.4, enabling GPU sharing in Kubernetes through a memory enforcement mechanism, allowing a single GPU to be split into hardware-isolated instances for multiple AI workloads, improving resource utilization. This aims to address the issue of many AI workloads only needing a portion of a GPU, enabling platform teams to serve more workloads with existing infrastructure. NVIDIA claims its GPUs already support various sharing methods. This feature allows a single GPU to be split into hardware-isolated instances for multiple AI workloads. NVIDIA emphasizes that many AI workloads only need a portion of a GPU, and sharing GPUs enables platform teams to serve more workloads with existing infrastructure.
AI-written from the reports · Updated 22 days ago
Coverage timeline
Follow the coverage to see every side of the story.
- NVIDIA Technical BlogMemory-Enforced GPU Sharing in Kubernetes with NVIDIA CUDA 13.4
NVIDIA CUDA 13.4 released, enabling GPU sharing in Kubernetes through memory-enforced mechanisms, allowing a single GPU to be split into hardware-isolated instances for multiple AI workloads, improving resource utilization.