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

Sparse matrix computation is crucial in various modern applications,including large-scale graph analytics, deep learning, and recommendationsystems. The performance of these algorithms varies greatly depending on thestructure of the input matrix, making it difficult to gain a comprehensiveunderstanding of sparse computation and its relationship to inputs, algorithms,and target machine architecture. Despite extensive research on certain sparsealgorithms, such as SpMV, the overall family of sparse algorithms has yet to beinvestigated as a whole. In this paper, we introduce SpChar, a workloadcharacterization methodology for general sparse computation. SpChar employstree-based models to identify the most relevant hardware and inputcharacteristics, starting from hardware and input-related metrics gathered fromPerformance Monitoring Counters and matrices. Our analysis enables the creationof a characterization loop that facilitates the optimization of sparsecomputation by mapping the impact of architectural features to inputs andalgorithmic choices. We apply SpChar to more than 600 matrices from theSuiteSparse Matrix collection and three state-of-the-art Arm CPUs to determinethe critical hardware and software characteristics that affect sparsecomputation. In our analysis, we determine that the biggest limiting factorsfor high-performance sparse computation are (1) the latency of the memorysystem, (2) the pipeline flush overhead resulting from branch misprediction,and (3) the poor reuse of cached elements. However, the degree to which thoseimpact the performance of a CPU greatly depends on the algorithm and the inputdata.

SpChar: Characterizing the Sparse Puzzle via Decision Trees