NVIDIA's Rubin HPC GPU Packs More Than 50,000 CUDA Cores, Ships in 2028
Summary
NVIDIA announces Rubin HPC, a GPU in its Vera Rubin platform with more than 50,000 CUDA cores and seventh-generation NVLink, built to speed the simulations agents use in science and engineering. Ian Buck, vice president of hyperscale and high-performance computing, says faster simulation means faster discovery. The company expects availability in 2028.
Key Points
- Rubin HPC adds a sparse data acceleration capability that keeps its record CUDA core count supplied with data during simulations.
- NVIDIA pairs Rubin HPC with its CUDA-X libraries, which give agents optimized algorithm implementations for simulation and analytics.
- Ian Buck says the GPU's balanced design scales data movement and compute together instead of increasing compute alone.