August 11, 2015
Purdue University, like most big research institutions, relies on a constantly evolving cyber infrastructure to support its large and diverse community of resea Read more…
July 8, 2013
Infiniband carried a slight majority of the Top 500 share this year at ISC, a trend that Mellanox says will continue, both in HPC and beyond. We discussed IB's reach and efficiencies with the company's Gilad Shainer to better understand where Ethernet and Infinband are... Read more…
June 21, 2013
To wrap up ISC, we wanted to collect some general visualized trends from this year's rankings. In this visual feature, we've provided information on operating systems, key vendors, processor and interconnect technologies and more. While there are a million ways to analyze the... Read more…
June 20, 2013
China is making significant strides towards its goals to use supercomputing as the next big platform to spur economic growth. The country now makes up 13% of the Top 500 (with 51% from the United States). We talk with Inspur's Chief Scientist of HPC Applications about how the system will be used to forward the goals of the national... Read more…
June 18, 2013
Today at the International Supercomputing Conference in Leipzing, Germany, Jack Dongarra presented on a proposed benchmark that could carry a bit more weight than its older Linpack companion. The high performance conjugate gradient (HPCG) concept takes into account new architectures for new applications, while shedding the floating point.... Read more…
June 17, 2013
The Top 500 list of the world's fastest computers has just been announced. Not surprisingly, since it's been reported on prior to the official announcement, the Chinese Tianhe-2 system tops the list. And that is an understatement. We talk with Jack Dongarra, Horst Simon, Hans Meuer and others from the.... Read more…
May 28, 2013
When so many folks from the HPC community come at us with credible details about something as important as the next top system on the planet, it's hard to ignore. To quiet things down (and hopefully bring forth more information) we've published the consistent details about what we know from (very) credible sources.about this year's upcoming Top500 announcement. While unconfimed, we have.... Read more…
March 26, 2013
European oil and gas giant, Total, has looked to SGI again to supply a super that meets their modeling and simulation needs--but that is focused on power and cooling. The result, based on the SGI ICE X, should pull a top ten ranking on this year's Top 500 list--the most powerful commercial.... Read more…
Many organizations looking to meet their CAE HPC requirements focus on the HPC on-premises hardware or cloud options. But one surprise that many find is that the bulk of their HPC total cost of ownership (TCO) comes from the complexity of integrating HPC software with CAE applications and in perfectly orchestrating the many technologies to use the hardware and CAE licenses optimally.
This white paper discusses how TotalCAE can significantly reduce TCO by offering turnkey, on-premises HPC systems and public cloud HPC solutions specifically for CAE simulation workloads that include integrated technology and software. The solutions, which TotalCAE fully manages, have allowed its clients to deploy hybrid HPC environments that deliver significant savings of up to 80%, faster-running workflows, and peace of mind since their entire solution is managed by professionals well-versed in HPC, cloud, and CAE technologies.
Five Recommendations to Optimize Data Pipelines
When building AI systems at scale, managing the flow of data can make or break a business. The various stages of the AI data pipeline pose unique challenges that can disrupt or misdirect the flow of data, ultimately impacting the effectiveness of AI storage and systems.
With so many applications and diverse requirements for data types, management systems, workloads, and compliance regulations, these challenges are only amplified. Without a clear, continuous flow of data throughout the AI data lifecycle, AI models can perform poorly or even dangerously.
To ensure your AI systems are optimized, follow these five essential steps to eliminate bottlenecks and maximize efficiency.
© 2023 HPCwire. All Rights Reserved. A Tabor Communications Publication
HPCwire is a registered trademark of Tabor Communications, Inc. Use of this site is governed by our Terms of Use and Privacy Policy.
Reproduction in whole or in part in any form or medium without express written permission of Tabor Communications, Inc. is prohibited.