I know Nvidia contributes substantially to opensource but I was wondering if there was a centralized location to find which versions of software are required to enable utilization of the latest nvidia hardware functionality. For example I saw that there is a special nvidia tensorflow which is 10x optimized. How can I find the same for pytorch, and other open source libraries. For example NEON instruction support was added to snappy 1.1.10 and AES-NI support was added to openssl v1.1.1w
Related topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
NVIDIA Releases Updates to CUDA-X AI Libraries | 0 | 335 | August 21, 2022 | |
Possibility of intergrate direct into tensorflow workflow? | 1 | 499 | December 20, 2019 | |
NVIDIA Deep Learning SDK Now Available | 0 | 309 | August 21, 2022 | |
Support for versions other than 1.3.0 | 4 | 1008 | October 12, 2021 | |
GPU and AI learning question? | 0 | 384 | May 19, 2019 | |
How install tensorflow with GPU | 1 | 2321 | April 21, 2018 | |
CUDA 11.0 compatible with Tensorflow 2.2? | 0 | 1683 | July 13, 2020 | |
TensorFlow 2.0 with Tighter TensorRT Integration Now Available | 0 | 290 | August 21, 2022 | |
OpenCV and the latest OpticalFlow | 0 | 497 | November 25, 2019 | |
Update to tensorflow? | 2 | 416 | June 27, 2019 |