Numerical behavior of NVIDIA tensor cores

We explore the floating-point arithmetic implemented in the NVIDIA tensor cores, which are hardware accelerators for mixed-precision matrix multiplication available on the Volta, Turing, and Ampere microarchitectures. Using Volta V100, Turing T4, and Ampere A100 graphics cards, we determine what pre...

Descripción completa

Detalles Bibliográficos
Autores principales: Fasi, Massimiliano, Higham, Nicholas J., Mikaitis, Mantas, Pranesh, Srikara
Formato: Online Artículo Texto
Lenguaje:English
Publicado: PeerJ Inc. 2021
Materias:
Algorithms and Analysis of Algorithms
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7959640/
https://www.ncbi.nlm.nih.gov/pubmed/33816984
http://dx.doi.org/10.7717/peerj-cs.330

Internet

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7959640/
https://www.ncbi.nlm.nih.gov/pubmed/33816984
http://dx.doi.org/10.7717/peerj-cs.330

Ejemplares similares