The complexity of AI models is rapidly increasing to meet new challenges such as conversational AI. Training them requires tremendous computational power and scalability.

NVIDIA A100 Tensor Cores with Tensor Float (TF32) precision provide up to 20x more performance over NVIDIA Volta, requiring no code changes to do so and providing an additional 2x boost with automatic mixed precision and FP16. Scaling to thousands of A100 GPUs is possible in combination with NVIDIA® NVLink®, NVIDIA NVSwitch™, PCI Gen4, NVIDIA® Mellanox® InfiniBand®, and the NVIDIA Magnum IO™ SDK.

Training workloads like BERT can be solved at scale with 2,048 A100 GPUs in under a minute, setting a world record for solution time.

For the largest models with massive data tables such as Deep Learning Recommendation Models (DLRM), the A100 80 GB achieves up to 1.3 TB of unified memory per node and provides up to 3 times the throughput of the A100 40 GB.

NVIDIA's leadership in MLPerf has been solidified by multiple performance records in AI training benchmarks across the industry.

To unlock the next generation of discovery, scientists are looking at simulations to better understand the world around us.

NVIDIA A100 introduces Tensor Cores with twice the precision, representing the biggest leap in performance for HPC since the introduction of GPUs. Combined with 80GB of the fastest graphics memory, researchers can reduce a previously 10-hour, double-precision simulation on A100 to less than four hours. HPC applications can also leverage TF32, achieving up to 11 times higher throughput on dense single-precision matrix multiplication tasks.

For those HPC applications with the largest data sets, the additional memory of the A100 80 GB provides up to a 2x increase in throughput in Quantum Espresso, a materials simulation. The massive memory and unmatched storage bandwidth make the A100 80 GB the ideal platform for next-generation workloads.