Dynamically managed data for CPU-GPU architectures

  • Authors:

  • Thomas B. Jablin;James A. Jablin;Prakash Prabhu;Feng Liu;David I. August

  • Affiliations:

  • Princeton University, Princeton, New Jersey;Brown University, Providence, Rhode Island;Princeton University, Princeton, New Jersey;Princeton University, Princeton, New Jersey;Princeton University, Princeton, New Jersey

  • Venue:
  • Proceedings of the Tenth International Symposium on Code Generation and Optimization
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

GPUs are flexible parallel processors capable of accelerating real applications. To exploit them, programmers must ensure a consistent program state between the CPU and GPU memories by managing data. Manually managing data is tedious and error-prone. In prior work on automatic CPU-GPU data management, alias analysis quality limits performance, and type-inference quality limits applicability. This paper presents Dynamically Managed Data (DyManD), the first automatic system to manage complex and recursive data-structures without static analyses. By replacing static analyses with a dynamic run-time system, DyManD overcomes the performance limitations of alias analysis and enables management for complex and recursive data-structures. DyManD-enabled GPU parallelization matches the performance of prior work equipped with perfectly precise alias analysis for 27 programs and demonstrates improved applicability on programs not previously managed automatically.