English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This paper examines the design space trade-offs of DNNs accelerators aimingto achieve competitive performance and efficiency metrics for all fourcombinations of dense or sparse activation/weight tensors. To do so, wesystematically examine the overheads of supporting sparsity on top of anoptimized dense core. These overheads are modeled based on parameters thatindicate how a multiplier can borrow a nonzero operation from the neighboringmultipliers or future cycles. As a result of this exploration, we identify afew promising designs that perform better than prior work. Our findings suggestthat even the best design targeting dual sparsity yields a 20%-30% drop inpower efficiency when performing on single sparse models, i.e., those with onlysparse weight or sparse activation tensors. We found that one can reuseresources of the same core to maintain high performance and efficiency whenrunning single sparsity or dense models. We call this hybrid architectureGriffin. Griffin is 1.2, 3.0, 3.1, and 1.4X more power-efficient thanstate-of-the-art sparse architectures, for dense, weight-only sparse,activation-only sparse, and dual sparse models, respectively.

Griffin: Rethinking Sparse Optimization for Deep Learning Architectures