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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Sparse tensor computing is a core computational part of numerous applicationsin areas such as data science, graph processing, and scientific computing.Sparse tensors offer the potential of skipping unnecessary computations causedby zero values. In this paper, we propose a new strategy for extending row-wiseproduct sparse tensor accelerators. We propose a new processing element calledMaple that uses multiple multiply-accumulate (MAC) units to exploit localclusters of non-zero values to increase parallelism and reduce data movement.Maple works on the compressed sparse row (CSR) format and calculates onlynon-zero elements of the input matrices based on the sparsity pattern.Furthermore, we may employ Maple as a basic building block in a variety ofspatial tensor accelerators that operate based on a row-wise product approach.As a proof of concept, we utilize Maple in two reference accelerators: Extensorand Matraptor. Our experiments show that using Maple in Matraptor and Extensorachieves 50% and 60% energy benefit and 15% and 22% speedup over the baselinedesigns, respectively. Employing Maple also results in 5.9x and 15.5x smallerarea consumption in Matraptor and Extensor compared with the baselinestructures, respectively.

Maple: A Processing Element for Row-Wise Product Based Sparse Tensor Accelerators