skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Milestone M7 Report: Reducing Excess Data Movement Part 2

Abstract

This milestone evaluates techniques to measure and, if possible, reduce data movement across all levels of the memory hierarchy, focusing on CPU/GPU page level data movement and on intra-GPU memory hierarchy. We quantitatively evaluate the efficacy of the techniques in reducing data movement and measure how performance tracks data movement reduction. We study a small collection of benchmarks and proxy mini-apps that run on advanced pre-exascale GPUs and on the Accelsim GPU simulator. Our approach has two thrusts: to measure advanced data movement reduction directives and techniques on the newest available GPUs, and to evaluate our benchmark set on simulated GPUs configured with architectural refinements to reduce data movement. We primarily evaluated NVidia-based architectures due to the unavailability of AMD GPU hardware and tools until very recently.

Authors:
 [1];  [1];  [2];  [3];  [3];  [1]
  1. Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
  2. Sandia National Laboratories (SNL), Albuquerque, NM, and Livermore, CA (United States)
  3. Univ. of Texas at El Paso, TX (United States)
Publication Date:
Research Org.:
Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
Sponsoring Org.:
USDOE National Nuclear Security Administration (NNSA); USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
OSTI Identifier:
1828120
Report Number(s):
LLNL-TR-828299
1043795
DOE Contract Number:  
AC52-07NA27344; NA0003525
Resource Type:
Technical Report
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING

Citation Formats

Peng, Ivy B., Sarkar, Abhik, Voskuilen, Gwen R., Moore, Shirley V., Long, Rogelio, and Gokhale, Maya B. Milestone M7 Report: Reducing Excess Data Movement Part 2. United States: N. p., 2021. Web. doi:10.2172/1828120.
Peng, Ivy B., Sarkar, Abhik, Voskuilen, Gwen R., Moore, Shirley V., Long, Rogelio, & Gokhale, Maya B. Milestone M7 Report: Reducing Excess Data Movement Part 2. United States. https://doi.org/10.2172/1828120
Peng, Ivy B., Sarkar, Abhik, Voskuilen, Gwen R., Moore, Shirley V., Long, Rogelio, and Gokhale, Maya B. 2021. "Milestone M7 Report: Reducing Excess Data Movement Part 2". United States. https://doi.org/10.2172/1828120. https://www.osti.gov/servlets/purl/1828120.
@article{osti_1828120,
title = {Milestone M7 Report: Reducing Excess Data Movement Part 2},
author = {Peng, Ivy B. and Sarkar, Abhik and Voskuilen, Gwen R. and Moore, Shirley V. and Long, Rogelio and Gokhale, Maya B.},
abstractNote = {This milestone evaluates techniques to measure and, if possible, reduce data movement across all levels of the memory hierarchy, focusing on CPU/GPU page level data movement and on intra-GPU memory hierarchy. We quantitatively evaluate the efficacy of the techniques in reducing data movement and measure how performance tracks data movement reduction. We study a small collection of benchmarks and proxy mini-apps that run on advanced pre-exascale GPUs and on the Accelsim GPU simulator. Our approach has two thrusts: to measure advanced data movement reduction directives and techniques on the newest available GPUs, and to evaluate our benchmark set on simulated GPUs configured with architectural refinements to reduce data movement. We primarily evaluated NVidia-based architectures due to the unavailability of AMD GPU hardware and tools until very recently.},
doi = {10.2172/1828120},
url = {https://www.osti.gov/biblio/1828120}, journal = {},
number = ,
volume = ,
place = {United States},
year = {Mon Oct 25 00:00:00 EDT 2021},
month = {Mon Oct 25 00:00:00 EDT 2021}
}