GPUrdma: GPU-side library for high performance networking from GPU kernels.
Feras DaoudAmir WatedMark SilbersteinPublished in: ROSS@HPDC (2016)
Keyphrases
- graphics processing units
- real time
- graphics hardware
- parallel processing
- gpu implementation
- parallel implementation
- general purpose
- gpu accelerated
- graphics processors
- peer to peer
- heterogeneous computing
- parallel computation
- cluster of workstations
- compute intensive
- feature space
- highly parallel
- kernel learning
- parallel computing
- kernel function
- computation intensive
- e learning
- data sets