For example, the tests show at equivalent throughput rates today’s DGX A100 system delivers up to 4x the performance of the system that used V100 GPUs in the first round of MLPerf training tests. Meanwhile, the original DGX-1 system based on NVIDIA V100 can now deliver up to 2x higher performance thanks to the latest software optimizations.EE Times analyzes the results over here and compares them with Google's TPU.
These gains came in less than two years from innovations across the AI platform. Today’s NVIDIA A100 GPUs — coupled with software updates for CUDA-X libraries — power expanding clusters built with Mellanox HDR 200Gb/s InfiniBand networking.
NVIDIA releases Ampere A100 GPU machine learning performance
Posted on Friday, Jul 31 2020 @ 13:52 CEST by Thomas De Maesschalck