Enhancing Distributed Neural Network Training Through Node-Based Communications

S Moreno-Álvarez, ME Paoletti… - IEEE transactions on …, 2023 - ieeexplore.ieee.org
The amount of data needed to effectively train modern deep neural architectures has grown
significantly, leading to increased computational requirements. These intensive …

C-Lop: Accurate contention-based modeling of MPI concurrent communication

Z Wang, H Chen, W Cai, X Dong, X Zhang - Parallel Computing, 2022 - Elsevier
MPI communication optimization is a crucial stage to optimize high-performance
applications. As a formal analysis of MPI communication, the communication performance …

Extending -Lop to model MPI blocking primitives on shared memory

Z Wang, H Chen, X Dong, W Cai, Y Kang… - The Journal of …, 2022 - Springer
MPI communication optimization is essential for high-performance applications. The
communication performance models have made some achievements in improving the …