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

For the architecture community, reasonable simulation time is a strongrequirement in addition to performance data accuracy. However, emerging bigdata and AI workloads are too huge at binary size level and prohibitivelyexpensive to run on cycle-accurate simulators. The concept of data motif, whichis identified as a class of units of computation performed on initial orintermediate data, is the first step towards building proxy benchmark to mimicthe real-world big data and AI workloads. However, there is no practical way toconstruct a proxy benchmark based on the data motifs to help simulation-basedresearch. In this paper, we embark on a study to bridge the gap between datamotif and a practical proxy benchmark. We propose a data motif-based proxybenchmark generating methodology by means of machine learning method, whichcombine data motifs with different weights to mimic the big data and AIworkloads. Furthermore, we implement various data motifs using light-weightstacks and apply the methodology to five real-world workloads to construct asuite of proxy benchmarks, considering the data types, patterns, anddistributions. The evaluation results show that our proxy benchmarks shortenthe execution time by 100s times on real systems while maintaining the averagesystem and micro-architecture performance data accuracy above 90%, evenchanging the input data sets or cluster configurations. Moreover, the generatedproxy benchmarks reflect consistent performance trends across differentarchitectures. To facilitate the community, we will release the proxybenchmarks on the project homepage http://prof.ict.ac.cn/BigDataBench.

Data Motif-based Proxy Benchmarks for Big Data and AI Workloads