I thought FPGAs were more power efficient than GPUs for many ML applications for... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

anonymousDan on May 21, 2020 | parent | context | favorite | on: VexRiscv is a quadcore, Linux-capable RISC-V softc...

I thought FPGAs were more power efficient than GPUs for many ML applications for example (CNNs aside)?

londons_explore on May 21, 2020 [–]

For unusual ML architectures, like 1 bit precision, that might be true.

For anything using floating point maths, it isn't true.

anonymousDan on May 22, 2020 | [–]

What about this paper then, section 5.1.2, energy efficiency: https://dl.acm.org/doi/10.1145/2694344.2694347

Substantial gains in terms of performance per watt for FPGAs?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact