News

NVIDIA Unveils Rubin HPC GPU with Over 50,000 CUDA Cores at SC26

NVIDIA announces Rubin HPC, a GPU designed to accelerate simulations for science, engineering, and analytics with over 50,000 CUDA cores and advanced data movement.

NVIDIA Unveils Rubin HPC GPU with Over 50,000 CUDA Cores at SC26

At the SC26 conference, NVIDIA introduced the Rubin HPC GPU, a new addition to its Vera Rubin platform, aimed at accelerating simulations and data processing for science, engineering, and analytics workflows. The GPU is designed to support agent-driven exploration of candidate materials, molecules, and designs, enabling faster scientific discovery and industrial innovation.

Technical Advancements in Rubin HPC

The Rubin HPC GPU boasts a record-breaking number of CUDA cores, exceeding 50,000 on a single chip. This unprecedented core count is complemented by sparse data acceleration capabilities, which ensure efficient data feeding to the cores. Additionally, the GPU incorporates NVIDIA's seventh-generation NVLink interconnect technology, enhancing GPU-to-GPU connectivity for large-scale computational tasks.

According to Ian Buck, NVIDIA's vice president of hyperscale and high-performance computing, "For agents tackling science and engineering, faster simulation means faster discovery. Rubin HPC accelerates the simulations agents use to test ideas and guide their next steps, helping scientists and engineers explore more possibilities and bring breakthroughs closer."

Optimized for Scientific Computing

Rubin HPC is designed to address the computational challenges of workloads such as climate modeling, genomic analysis, engineering simulations, and financial risk assessment. These applications often involve retrieving scattered values from massive datasets, performing calculations, and repeating the process at scale. By balancing data movement and computation, Rubin HPC ensures that its CUDA cores spend more time computing rather than idling.

The GPU also integrates NVIDIA's CUDA-X libraries, which provide optimized implementations of complex algorithms commonly used in simulations and analytics. This integration allows agents to achieve faster reasoning outcomes, reducing the time required for hypothesis testing, simulation, and refinement.

Availability and Future Impact

Rubin HPC is expected to become available in 2028. Its balanced design and high computational power are anticipated to accelerate breakthroughs across industries, from scientific research to industrial design. By shortening the feedback loop between hypothesis and discovery, the GPU aims to redefine the pace of innovation in agent-driven workflows.

Sign in to save this article Free account: build your research collection and return to read it anytime.