GTC'14 – OpenACC speeds up!
By   |  March 27, 2014

What a more appropriate place than GTC to announce good news regarding OpenACC? The first one is the addition of a new member, oil and gas giant Total, whose teams are at the forefront of supercomputing in the seismic field. For the record, Total was one of the first private companies to adopt directives programming, and more specifically, CAPS’ HMPP model, as early as 2007.

Then came the announcement of the creation of a new benchmark suite by SPEC’s High Performance Group, SPEC V1.0 ACCEL, which includes a series of intensive parallel applications to measure the performance of hybrid systems.

Last, OpenACC is starting a new program to reward its most valuable contributors. The first winner will also be announced during the conference, along with over twenty topical tutorials and other sessions relating to compiler directives for accelerated computing, portability and performance on different platforms, OpenACC parallelization and the performance of the NIM climate model on Titan. We’ll closely follow these events, and for the same price, prepare a long interview with Duncan Poole, OpenACC’s President.

These news come in all the more timely as the consortium, for the most part originally supported by NVIDIA, clearly intends to establish itself as a de facto standard, particularly against its rival OpenMP. They must also be taken in the context of OpenMP’s last month release of a new roadmap or mission statement that formalizes its opening to accelerators along with shared memory systems. As Henry Fielding once wrote, worth begets in base minds, envy; in great souls, emulation.

© HPC Today 2024 - All rights reserved.

Thank you for reading HPC Today.

Express poll

Do you use multi-screen
visualization technologies?

Industry news

Brands / Products index