TSPLIT: Fine-grained GPU Memory Management for Efficient DNN Training via Tensor Splitting.

Xiaonan Nie Xupeng Miao Zhi Yang Bin Cui

Published in: ICDE (2022)

Keyphrases

fine grained
coarse grained
memory management
access control
training process
operating system
real time
higher order
tightly coupled
massively parallel
parallel computation
database
machine learning
concurrency control
parallel implementation