The Week in HPC News
This was anything but the quiet post-International Supercomputing Conference and Top 500 announcement week we were expecting with a slew of interesting news announcements, topped of course, by the much-awaited details of what we can expect with one of the United States’ largest upcoming systems (that we know about, anyway), the Trinity machine.
The initial news filtered out of Cray this morning, sending their stock price leaping—and us into rush mode to find out more about the system. The first announcement offered nothing in terms of system size, other than noting that it would be a next-generation Cray XC supercomputer, decked out with Haswell and Knight’s Landing. We were able to secure the only known details following a conversation with a lead on the joint effort between Los Alamos and Sandia. You can read more about that here.
It seemed absurd that just earlier this very week, we published a speculative piece about Cray’s role in the coming year and its observations about the market. If only we’d held onto to that analysis for another day. But moving on…
All of the noise around the U.S. supercomputer pushed out a smaller, less confirmed item (we’re still working on this) about a coming upgrade to the existing #1 Tianhe-2 machine in China. Word from a Chinese source indicates that progress began today to give the system a 100 petaflop upgrade which, while not out of the realm of possibility, still needs to be verified. The source says that beginning this week at the National Supercomputer Center in Guangzhou, a hardware upgrade is beginning—one that will be powered by native Chinese processors. This 1000 petaflop upgrade is expected to be complete by 2015.
These are all definitely top-ranking systems (even if they don’t run Linpack), which complements the podcast topic we carried this week. Among our conversations, we spoke with Dr. Michael Heroux from Sandia about the coming HPCG benchmark to better understand the results published here. If you have a few extra minutes this weekend, it’s definitely worth a listen.
Stay tuned next week for some great stuff out of the XSEDE conference, an interview with IDC’s Research VP, Steve Conway about the HPC market this coming year, new work being done at IBM Research, and more podcast interviews.
Aside from these news items, let’s take a look at what else transpired in this surprisingly active week in high performance computing.
More Top News Items This Week
IBM has announced it is investing $3 billion over the next 5 years in two broad research and early stage development programs to push the limits of chip technology needed to meet the emerging demands of next-generation systems. The first research program is aimed at so-called “7 nanometer and beyond” silicon technology that will address serious physical challenges that are threatening current semiconductor scaling techniques and will impede the ability to manufacture such chips. The second is focused on developing alternative technologies for post-silicon era chips using entirely different approaches, which IBM scientists and other experts say are required because of the physical limitations of silicon based semiconductors.
EZchip is acquiring Tilera, makers of multi-core processors, intelligent network interface cards and white-box appliances for data-center networking equipment. Under the terms of the agreement, EZchip will pay Tilera’s stockholders up to $130 million in cash; of which $50 million is payable at closing and up to an additional $80 million is payable subject to the attainment of certain future performance milestones. The transaction is expected to close in the third quarter of 2014.
Intel has chosen the Hartree Centre, part of the Science and Technology Facilities Council (STFC), as its newest Intel Parallel Computing Center. This new collaboration allows Intel and STFC’s Hartree Centre to work together on designing and testing software for the supercomputers of the future.
A new contract from the High Performance Computing Modernization Program Defense Research Engineering Network contract for ASRC Federal InuTeq has been signed. The five-year contract, consisting of a base year plus four annual options, has a maximum value of nearly $88 million. ASRC Federal InuTeq will provide the High Performance Computing Modernization Program with technical and professional support elements required for the operation. Specifically, the team’s network engineers and subject matter experts will focus on running DREN operation centers, ensuring the security, availability and reliability of the network.
We also carried a few user stories of note this week, including one about the role supercomputers are playing in health research in the U.K. as well as how scientific users are leveraging the Stampede system at TACC. Both are worth a read…
On the Road
And don’t forget to look for news from the XSEDE conference in Atlanta this coming week.