NVbandwidth is a CUDA-based tool that measures bandwidth and latency for various memory copy patterns across different links using either copy engine (CE) or kernel copy methods. It reports the current measured bandwidth on your system, providing valuable insights into the performance characteristics of your GPU setup. While modern GPUs boast impressive compute capabilities, their performance is frequently limited by how quickly data can be moved between different devices: CPU memory to GPU memory GPU memory to CPU memory GPU memory to GPU memory Understanding these performance characteristi
NVIDIA NVbandwidth: Your Essential Tool for Measuring GPU Interconnect and Memory Performance | NVIDIA Technical Blog…View all posts by Phoebe Lee View all posts by Phoebe Lee About Eric Kana Eric Kana is an NVIDIA principal solution architect with a focus on GPU virtualization for both datacenter…
…View all posts by Aiden Chang View all posts by Aiden Chang About Amparo Canaveras Amparo Canaveras is a senior solutions architect at NVIDIA, specializing in generative AI applications within the telecommunications…
…View all posts by Max Bickley View all posts by Max Bickley About William Cannady Will Cannady is a technical product manager with the NVIDIA Omniverse team. He works closely with partners…
…View all posts by Christian Shrauder View all posts by Christian Shrauder About Gail Frederick Gail Frederick leads engineering for capacity automation and trust services at NVIDIA DGX Cloud. Her teams deliver…
…View all posts by Chris Alexiuk View all posts by Chris Alexiuk About Isabel Hulseman Isabel Hulseman is a product marketing manager at NVIDIA, where she focuses on enterprise AI software and…
…View all posts by Ashley Goldstein View all posts by Ashley Goldstein About Brian Harrison Brian Harrison is senior director of software product management of digital twins for Omniverse. He’s been…
…This post provides a walkthrough of key UST features, implementation details, and performance overview, including: Zero-cost interoperability: Data-movement-free conversion with PyTorch, SciPy, and CuPy. Custom formats: Define novel sparsity…
…LLM team’s post-training job waits…and waits…and waits. The LLM team’s inference services are fine, and the guaranteed quota works perfectly. But their post-training job is effectively…
…View all posts by Lucas Liebenwein View all posts by Lucas Liebenwein About Suyog Gupta Suyog Gupta is a distinguished engineer and manager at NVIDIA where he works on inference software architecture…
…View all posts by Anton Polyakov View all posts by Anton Polyakov About Fagani Hajizada Fagani Hajizada is a senior software engineer at NVIDIA, focused on scheduling and orchestration for HPC and…