GPU/AI Computing
News
Unlocking Tensor Core Performance with Floating Point Emulation in cuBLAS
Unlocking Tensor Core Performance with Floating Point Emulation in cuBLAS
<img alt="Decorative image." class="webfeedsFeaturedVisual wp-post-image" height="431" src="https://developer-blogs.nvidia.com/wp-content/uploads/2025/10/floating-cubes-768x431-png.webp" style="displa...
NVIDIA CUDA-X math libraries provide the fundamental numerical building blocks that enable developers to deploy accelerated applications across multiple...NVIDIA CUDA-X math libraries provide the fundamental numerical building blocks that enable developers to deploy accelerated applications across multiple high-performance domains, including AI and scientific computing. cuBLAS is a CUDA-X math library that consists of a highly optimized collection of basic linear algebra subroutines for matrix and vector operations that are specifically tuned…
Source: NVIDIA Technical Blog
Word count: 1168 words
Published on 2025-10-25 00:21