Data placement on GPUs
X. Shen1; B. Wu2 1 North Carolina State University, Raleigh, NC, United States2 Colorado School of Mines, Golden, CO, United States
Abstract
Modern graphics processing unit (GPU) memory systems consist of a number of components with different properties. How data are placed on the various memory components often causes a significant difference in a program’s performance. This chapter discusses the complexity of GPU memory systems and describes a software framework named PORPLE to show how to automatically resolve the complexity for a given GPU program.
Keywords
Cache; Compiler optimization; Data placement; Hardware specification language; GPUs; Memory systems
1 Introduction
To meet the relentless demands ...
Get Advances in GPU Research and Practice now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
