Victor Hugo Rosales

This paper discusses how to decrease and stabilize network latency in a Beowulf system. Having low latency is particularly important to reduce execution time of High Performance Computing applications. Optimization opportunities are identified and analyzed over the different system components that are integrated in compute nodes, including device drivers, operating system services and kernel parameters. This work contributes with a systematic approach to optimize communication latency, provided… Mostrar más

This paper discusses how to decrease and stabilize network latency in a Beowulf system. Having low latency is particularly important to reduce execution time of High Performance Computing applications. Optimization opportunities are identified and analyzed over the different system components that are integrated in compute nodes, including device drivers, operating system services and kernel parameters. This work contributes with a systematic approach to optimize communication latency, provided with a detailed checklist and procedure. Performance impacts are shown through the figures of benchmarks and mpiBLAST as a real-world application. We found that after applying different techniques the default Gigabit Ethernet latency can be reduced from about 50 μs into nearly 20 μs. Mostrar menos

In order to take advantage of the multi-core architecture of modern processors, the legacy serial code must be analyzed to discover the regions where the parallelization effort can be more rewarding. This paper reports the parallelization and optimization procedures done on an HPC application. The Intel R VTune Amplifier XE tool was used to profile the NEC2C software for simulation of electromagnetic response of antennas, and the proposed parallelization was implemented. The application got a… Mostrar más

In order to take advantage of the multi-core architecture of modern processors, the legacy serial code must be analyzed to discover the regions where the parallelization effort can be more rewarding. This paper reports the parallelization and optimization procedures done on an HPC application. The Intel R VTune Amplifier XE tool was used to profile the NEC2C software for simulation of electromagnetic response of antennas, and the proposed parallelization was implemented. The application got a speed up of nearly 7 times before a saturation effect was reached. Mostrar menos

In order to take advantage of the multi-core architecture of modern processors, the legacy serial code must be analyzed to discover the regions where the parallelization effort can be more rewarding. This paper reports the parallelization and optimization procedures done on an HPC application. The Intel R VTune Amplifier XE tool was used to profile the NEC2C software for simulation of electromagnetic response of antennas, and the proposed parallelization was implemented. The application got a… Mostrar más

In order to take advantage of the multi-core architecture of modern processors, the legacy serial code must be analyzed to discover the regions where the parallelization effort can be more rewarding. This paper reports the parallelization and optimization procedures done on an HPC application. The Intel R VTune Amplifier XE tool was used to profile the NEC2C software for simulation of electromagnetic response of antennas, and the proposed parallelization was implemented. The application got a speed up of nearly 7 times before a saturation effect was reached. Mostrar menos

This is a brief introduction to the Yocto build system

This paper discusses how to decrease and stabilize network latency in a Beowulf system. Having low latency is particularly important to reduce execution time of High Performance Computing applications. Optimization opportunities are identified and analyzed over the different system components that are integrated in compute nodes, including device drivers, operating system services and kernel parameters. This work contributes with a systematic approach to optimize communication latency, provided… Mostrar más

This paper discusses how to decrease and stabilize network latency in a Beowulf system. Having low latency is particularly important to reduce execution time of High Performance Computing applications. Optimization opportunities are identified and analyzed over the different system components that are integrated in compute nodes, including device drivers, operating system services and kernel parameters. This work contributes with a systematic approach to optimize communication latency, provided with a detailed checklist and procedure. Performance impacts are shown through the figures of benchmarks and mpiBLAST as a real-world application. We found that after applying different techniques the default Gigabit Ethernet latency can be reduced from about 50 μs into nearly 20 μs. Mostrar menos

Prepared a theoretical and practical course of distributed programming with MPI 5 days course for high level students. The course covered from an actual installation and deployment of a real cluster to actually code some exercises using MPI and running them in the cluster thru a resource manager.
I dictated that course together with a couple of colleagues for over 80 high level students which had to deliver a final exam also prepared by me and my colleagues.