This document describes the work carried out by the COKA project in 2014. In summary, the project has focused on the use of MIC-based accelerator boards as working horses for several computational applications in theoretical physics. These accelerators were released by Intel at the beginning of 2013 under the name Xeon-Phi. We have worked on the development of the best optimization strategies to use these processor efficiently, including an analysis of the available programming environments, and have assessed the performance and usability of these systems with respect to more standard processors. We have also started to consider the best options to use parallel systems containing several MIC accelerators.
The COKA project / Alfieri, Roberto; Brambilla, Michele; Calore, Enrico; DE PIETRI, Roberto; DI RENZO, Francesco; Feo, Alessandra; Schifano Sebastiano, Fabio; Tripiccione, Raffaele. - In: INFN-CNAF ANNUAL REPORT. - ISSN 2283-5490. - (2015), pp. 121-127.
The COKA project
ALFIERI, Roberto;BRAMBILLA, Michele;DE PIETRI, Roberto;DI RENZO, Francesco;FEO, Alessandra;
2015-01-01
Abstract
This document describes the work carried out by the COKA project in 2014. In summary, the project has focused on the use of MIC-based accelerator boards as working horses for several computational applications in theoretical physics. These accelerators were released by Intel at the beginning of 2013 under the name Xeon-Phi. We have worked on the development of the best optimization strategies to use these processor efficiently, including an analysis of the available programming environments, and have assessed the performance and usability of these systems with respect to more standard processors. We have also started to consider the best options to use parallel systems containing several MIC accelerators.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
