Prosperity : Accelerating Spiking Neural Networks via Product Sparsity

Chiyue Wei † \dagger , Cong Guo † \dagger 1 , Feng Cheng † \dagger , Shiyu Li † \dagger , Hao (Frank) Yang ‡ \ddagger , Hai (Helen) Li † \dagger , Yiran Chen † \dagger · 2025

Read Prosperity : Accelerating Spiking Neural Networks via Product Sparsity by Chiyue Wei † \dagger , Cong Guo † \dagger 1 , Feng Cheng † \dagger , Shiyu Li † \dagger , Hao (Frank) Yang ‡ \ddagger , Hai (Helen) Li † \dagger , Yiran Chen † \dagger (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-03379.txt.

Description

Chiyue Wei and six co-authors teach spiking neural networks to skip repeated arithmetic by spotting repeated products. Spiking nets are frugal because most activations are zero, and chips already skip those zeros. Yet many nonzero inner products repeat across a matrix multiply, and bit sparsity cannot see them. Their product sparsity idea caches and reuses those repeated products, cutting real work.

On a SpikeBERT model the trick dropped compute density to 1.23% and cut operations elevenfold against bit sparsity. A chip design named Prosperity finds the repeats on the fly. It ran 7.4 times faster than a prior accelerator and far leaner on energy than a flagship GPU.

Catalog record

CREATORChiyue Wei † \dagger , Cong Guo † \dagger 1 , Feng Cheng † \dagger , Shiyu Li † \dagger , Hao (Frank) Yang ‡ \ddagger , Hai (Helen) Li † \dagger , Yiran Chen † \dagger
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE65 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-03379

Browse the full library · About the Tower of Babel