view article
Figure 3
The GPU processing pipeline from top to bottom: DMA transfer, CPU to GPU transfer, data processing, GPU to CPU transfer. The pipeline involves three heterogeneous systems: RNIC, CPU and GPU. The RNIC cannot directly synchronize the GPU. Each stage in the execution pipeline is synchronized by an RDMA event or conditionally by the veto kernel result.

Journal logoJOURNAL OF
SYNCHROTRON
RADIATION
ISSN: 1600-5775
Volume 27| Part 5| September 2020| Pages 1297-1306
https://doi.org/10.1107/S1600577520008140
Follow J. Synchrotron Rad.
Sign up for e-alerts
E-alerts
Follow J. Synchrotron Rad. on Twitter
Twitter
Follow us on facebook
Facebook
Sign up for RSS feeds
RSS