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

Modern deep learning models have high memory and computation cost. To makethem fast and memory-cost efficient, structured model pruning is commonly used.We find that pruning a model using a common training accelerator with largesystolic arrays is extremely performance-inefficient. To make a systolic arrayefficient for pruning and training, we propose FlexSA, a flexible systolicarray architecture. FlexSA dynamically reconfigures the systolic arraystructure and offers multiple sub-systolic operating modes, which are designedfor energy- and memory bandwidth-efficient processing of tensors with differentsizes and shapes. We also present a compilation heuristic for tilingmatrix-multiplication-and-accumulation operations in a training workload tobest utilize the resources of FlexSA. Based on our evaluation, FlexSA with theproposed compilation heuristic improves compute resource utilization of pruningand training modern CNN models by 37% compared to a conventional trainingaccelerator with a large systolic array. FlexSA also improves on-chip datareuse by 1.7X saving 28% energy compared to naive systolic array splitting.

FlexSA: Flexible Systolic Array Architecture for Efficient Pruned DNN Model Training