J
Juan C. Díaz-Martín
Researcher at University of Extremadura
Publications - 9
Citations - 58
Juan C. Díaz-Martín is an academic researcher from University of Extremadura. The author has contributed to research in topics: Computational resource & Shared memory. The author has an hindex of 3, co-authored 9 publications receiving 36 citations.
Papers
More filters
Journal ArticleDOI
A Survey of Communication Performance Models for High-Performance Computing
TL;DR: In this article, the authors present the state-of-the-art in analytic communication performance models, providing sufficiently detailed descriptions of particularly noteworthy efforts, as well as future directions for future research in the area of analytical communication performance modeling.
Journal ArticleDOI
Training deep neural networks: a static load balancing approach
Sergio Moreno-Álvarez,Juan M. Haut,Mercedes E. Paoletti,Juan A. Rico-Gallego,Juan C. Díaz-Martín,Javier Plaza +5 more
TL;DR: This work studies the application of well-known HPC static load balancing techniques to the distributed training of deep models, and shows that, while the classification accuracy is kept constant, the training time substantially decreases with respect to unbalanced training.
Proceedings ArticleDOI
Formal modeling and performance evaluation of a run-time rank remapping technique in Broadcast, Allgather and Allreduce MPI collective operations
TL;DR: An alternate approach is works out, departing from the a priori knowledge of a regular mapping as round-robin or sequential, and keeping the original algorithm unmodified, switches at run time the rank to process mapping into another regular mapping that reduces network traffic.
Proceedings ArticleDOI
A Pthreads-Based MPI-1 Implementation for MMU-Less Machines
Juan A. Rico-Gallego,Jesus M. Alvarez-Llorente,F.J. Perogil-Duque,P.P. Antunez-Gomez,Juan C. Díaz-Martín +4 more
TL;DR: Design issues are discussed, memory footprint and performance figures in a Xilkernel/PowerPC405 Xilinx FPGA are given and AzequiaMPI is described.
Journal ArticleDOI
A tool to assess the communication cost of parallel kernels on heterogeneous platforms
TL;DR: The utility of the $$\uptau$$ τ –Lop analytic communication performance model in facing issues is discussed and a practical symbolic computation tool that represents, manipulates and accurately evaluates the formal communication cost expression derived from a hybrid kernel is contributed.