NVIDIA CUDA 10 Adding nvJPEG, Turing Support
Written by Michael Larabel in NVIDIA on 14 August 2018 at 05:19 AM EDT. 8 Comments
NVIDIA is capitalizing upon SIGGRAPH 2018 as now in addition to launching the Quadro RTX GPUs and open-sourcing the MDL SDK they have announced their work on CUDA 10.

CUDA 10 is being geared for release with many improvements over CUDA 9 for building GPU-accelerated applications with this proprietary API alternative to OpenCL.

Most notable to CUDA 10 is initial support for Turing GPUs, like the just-announced Quadro RTX series and the upcoming GeForce GTX "2000" series or whatever they end up being labeled.

CUDA 10 also has more of the never-ending optimization work on their libraries for delivering optimal performance, integrates nvJPEG support for faster JPEG image processing, CUDA Graphs, new interoperability features between CUDA and Vulkan/D3D12, and various new developer tools. The library performance tuning work for CUDA 10 specifically calls out stronger FFT performance, dense linear algebra routines, and mixed-precision GEMM performance.

More details on the forthcoming CUDA 10 via developer.nvidia.com.
Related News
About The Author
Author picture

Michael Larabel is the principal author of Phoronix.com and founded the site in 2004 with a focus on enriching the Linux hardware experience. Michael has written more than 20,000 articles covering the state of Linux hardware support, Linux performance, graphics drivers, and other topics. Michael is also the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org automated benchmarking software. He can be followed via Twitter or contacted via MichaelLarabel.com.

Popular News This Week