Google Cloud users now have access to NVIDIA’s Ampere GPUs to complete their AI workloads. Early this week, the two tech giants announced the introduction of NVIDIA A100 Tensor Core which was launched just last month. NVIDIA claims that A100 “has come to the cloud faster than any NVIDIA GPU in history.”

The addition of A100 will have a lot better performance, increased memory, highly flexible precision support, and increased process isolation for running multiple workloads on one GPU which means there are several benefits for Google Cloud users who have ML and HPC workloads. 

With A100, Google Cloud users achieve new capabilities such as AI training and inference, data analytics, scientific computing, genomics, edge video analytics, 5G services, and more.

As compared to its predecessors, NVIDIA’s A100 packs a punch - according to the company, A100 augments performance related to training and interference by 20x as compared to previous GPUs. Models like Google’s BERT, an AI language model, can be training in under 40 minutes by using 1024 A100s! 

“Google Cloud customers often look to us to provide the latest hardware and software services to help them drive innovation on AI and scientific computing workloads.

With our new A2 VM family, we are proud to be the first major cloud provider to market NVIDIA A100 GPUs, just as we were with NVIDIA T4 GPUs. We are excited to see what our customers will do with these new capabilities,” said Manish Sainani, Director of Product Management at Google Cloud. 

The new Ampere-based GPUs are now available on Google Cloud in Alpha. Users can access up to 16 A100 GPUs, which is a total of 640GB of GPU memory and 1.3TB of system memory.