Back to news

AI · Official

NVIDIA Speeds Up Scientific Simulations With New cuBLAS Algorithms

NVIDIA Technical blog ·

At a glance

NVIDIA added new algorithms to its cuBLAS library that speed up the high-precision matrix multiplication used in scientific simulations. The results keep their full accuracy. The change is available in CUDA Toolkit 13.4.2.

Key points

Why it matters

Scientific simulations, such as those for climate, new materials, and drug discovery, spend most of their time on matrix multiplication. Speeding up this step without losing accuracy can shorten research time and lower costs.

Original source

Faster Scientific Simulations with New FP64 Algorithms in NVIDIA cuBLAS

NVIDIA Technical blog · Nov 14, 07:13 KST

Read on NVIDIA Technical blog ↗

AI summarized the original article. Check the original for accuracy. Copyright belongs to NVIDIA Technical blog.

AI