NVIDIA A100 Tensor Core GPU: Performance and Innovation

By Jack Choquette, Wishwesh Gandhi, Olivier Giroux, Nick Stam, Ronny Krashinsky. IEEE Micro, 2021

425 citations recorded by the source corpus.

NVIDIA A100 Tensor Core GPU is NVIDIA's latest flagship GPU. It has been designed with many new innovative features to provide performance and capabilities for HPC, AI, and data analytics workloads. Feature enhancements include a Third-Generation Tensor Core, new asynchronous data movement and programming model, enhanced L2 cache, HBM2 DRAM, and third-generation NVIDIA NVLink I/O.

Subfields: Computer Networks and Communications, Hardware and Architecture

View Paper: https://www.semanticscholar.org/paper/21d0613c3e7fe2cb31f34441c1604edc9882fa45

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

1. Daniel Sánchez
2. Srini Devadas
3. Chris Peikert
4. Karim M. El Defrawy
5. Vadim Lyubashevsky
6. Jintai Ding
7. Oded Regev
8. Dan Boneh

In Hardware and Architecture

1. Daniel Sánchez
2. Srini Devadas
3. Karim M. El Defrawy
4. Roel Maes
5. Patrick Schaumont
6. Abhranil Maiti
7. G. Suh

Citations in the corpus

• The operational medium-range deterministic weather forecasting can be extended beyond a 10-day lead time (2023, neutral): “4KW for an A100 (Choquette et al., 2021), a 10-day inference by FengWu consumes”
• F1: A Fast and Programmable Accelerator for Fully Homomorphic Encryption (2021, methodology): “We assume 512GB/s bandwidth per PHY; this is similar to the NVIDIA A100 GPU [18], which has 2.”