NVIDIA last month announced a record-breaking benchmark result of 410 trillion traversed edges per second, ranking No. 1 on the 31st Graph500 breadth-first search list.

The NVIDIA Developer Blog provides developers with a  knowledge on GPU computing, AI, and deep learning, offering tutorials, code samples, and real-world applications of NVIDIA technologies. From optimizing GPU-accelerated algorithms to implementing  AI models, developers can learn practical techniques and strategies for harnessing the power of NVIDIA GPUs in their projects. Moreover, the blog highlights advancements in GPU architectures, CUDA programming, and GPU-accelerated libraries, empowering developers to stay at the forefront of GPU computing innovation.

NVIDIA

NVIDIA achieved a record-breaking Graph500 benchmark result of 410 trillion traversed edges per second using 8,192 H100 GPUs on CoreWeave's cloud infrastructure. The breakthrough came from reengineering graph processing to run entirely on GPUs using custom active messaging software (IBGDA and NVSHMEM), bypassing CPUs. This approach delivered 3x better performance per dollar compared to traditional CPU-based systems, using just over 1,000 nodes versus 9,000 nodes for comparable results. The technology enables processing massive sparse, irregular workloads and has implications for HPC fields like fluid dynamics and weather forecasting.

How NVIDIA H100 GPUs on CoreWeave’s AI Cloud Platform Delivered a Record-Breaking Graph500 Run

Reengineering Graph Processing for the GPU