Open access
Jul 2026
DGNA: Dissecting GPU NUMA Architecture through Microbenchmarking and Data Analysis
This paper introduces DGNA, a methodology designed to unveil the NUMA architecture of the GPU memory hierarchy through microbenchmarking and data analysis, and is the first paper to detail the NUMA architecture within the GPU memory subsystem.
Changxi Liu, Yun Chen, Trevor E. Carlson
· ACM Transactions on Architec... · 1 citation