Milestone M7 Report: Reducing Excess Data Movement Part 2

Technical Report · Mon Oct 25 00:00:00 EDT 2021

DOI:https://doi.org/10.2172/1828120· OSTI ID:1828120

Peng, Ivy B. ^[1]; Sarkar, Abhik ^[1]; Voskuilen, Gwen R. ^[2]; Moore, Shirley V. ^[3]; Long, Rogelio ^[3]; Gokhale, Maya B. ^[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)

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.

• Research Organization: Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
• Sponsoring Organization: USDOE National Nuclear Security Administration (NNSA); USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
• DOE Contract Number: AC52-07NA27344; NA0003525
• OSTI ID: 1828120
• Report Number(s): LLNL-TR-828299; 1043795
• Country of Publication: United States
• Language: English

Related Subjects

97 MATHEMATICS AND COMPUTING