Enhancing Sparse Data Processing through Energy-Efficient Heterogeneous Architectures

Efficiently processing sparse data is a significant challenge in many high-impact domains, including machine learning, graph analytics, and scientific computing. Sparse data representations, such as compressed sparse row (CSR), often suffer from substantial overheads during indexing and irregular memory access patterns, limiting the performance and scalability of these applications. This research addresses these challenges by exploring innovative hardware solutions to improve the performance and energy efficiency of sparse linear algebra computations. By leveraging specialized ASIC accelerators and RISC cores as processing-in-memory (PIM) units, and optimizing parallel data processing on GPUs, this work seeks to significantly reduce computational bottlenecks and enhance the capabilities of applications that rely on sparse data.

Identiferoai:union.ndltd.org:unt.edu/info:ark/67531/metadc2415970
Date12 1900
CreatorsVasireddy, Pranathi
ContributorsZhao, Hui, Kavi, Krishna M, Mehta, Gayatri, Ji, Yuede, Shelor, Charles
PublisherUniversity of North Texas
Source SetsUniversity of North Texas
LanguageEnglish
Detected LanguageEnglish
TypeThesis or Dissertation
FormatText
RightsPublic, Vasireddy, Pranathi, Copyright, Copyright is held by the author, unless otherwise noted. All rights Reserved.

Page generated in 0.002 seconds