Here’s a collection of highlights, selected totally subjectively, from this week’s HPC news stream as reported at insideHPC.com and HPCwire.
>>10 words and a link
LNXI holds fire sale, everything must go;
http://insidehpc.com/2008/04/29/lnxi-holds-a-fire-sale/
A new numerical libraries group at MS focused on HPC;
http://insidehpc.com/2008/05/01/a-new-numerical-libraries-group-at-ms-focused-on-hpc/
Cray reports Q1 loss, optimistic for rest of 2008;
http://insidehpc.com/2008/04/29/cray-reports-q1-results-posts-loss-2/
Square Kilometre Array project to grow HPC in South Africa;
http://insidehpc.com/2008/04/27/expanding-south-africas-supercomputing-infrastructure-in-a-big-way/
Microsoft funds four centers for green computing research;
http://insidehpc.com/2008/04/29/microsoft-funds-green-computing-research/
Green HPC panel at Platform Global in May;
http://insidehpc.com/2008/05/01/green-hpc/
Max Planck Institute builds Top50 machine with Woven ethernet inside;
http://insidehpc.com/2008/04/29/good-news-for-woven/
HPC pr0n: Video tour of TACC’s ranger installation;
http://blogs.sun.com/jonasdias/entry/tacc_ranger_tour
>>Cray and Intel collaboration
Cray and Intel announced this week that they have inked a multi-year agreement to advance high performance computing on Intel microprocessors. What what what!? Cray will deliver these advancements on their future supercomputing platforms.
“We’re excited at the potential of bringing together Intel’s powerful silicon expertise and Cray’s industry leadership in scalable HPC systems,” said Peter Ungaro, president and CEO of Cray. “We pride ourselves in offering the most innovative supercomputing systems and our customers will now enjoy greater choice in processor technologies.”
The two companies plan to dive well into multicore processor architectures and next-generation interconnects. This will result in a wide range of HPC systems and technologies over the next several years.
“Cray’s commitment to Intel is a testament of our commitment to HPC and the strength of our hardware and software roadmap and many-core research,” said Patrick Gelsinger, senior vice president and general manager of Intel’s Digital Enterprise Group. “Throughout Cray’s history, it has been an innovator in high-end HPC while Intel has pushed the boundaries of processor technology.”
This is certainly a departure from Cray’s previous stance of AMD-centric compute architectures. This also becomes a threat to SGI’s monopoly of Intel silicon in big-iron compute.
Read the full release here, and find more analysis and coverage at HPCwire here and here.
>>It sucks to be AMD
Brooke Crothers over at CNET Blogs posted yesterday about another pearl in AMD’s bag of sorrow, this time caused by some bad communication with manufacturers using their new Phenom chip:
AMD confirmed Monday that some motherboard suppliers are mismatching high-end quad-core Phenom processors with a lower-end chipset.
…
“They’ve taken an enthusiast-class quad-core part and paired it with a mainstream motherboard,” Whitman said. “And not all motherboard manufacturers have tweaked their boards to support a 125-watt TDP [Thermal Design Power].” Whitman says that AMD’s 790 chipset — not the 780 — should be paired with the 9750 and 9850 processors and that a number of motherboard makers are already doing this.
I’ll grant that this one doesn’t appear to the be the fault of AMD’s engineering or manufacturing teams, but certainly some component of AMD’s channel communications program appears to be broken. They just can’t catch a break.
>>New fundamental circuit element discovered
The DailyTech reported on an innovation out of HP labs that appeared in a Nature article earlier this week.
Researchers at HP Labs, the central research center for the company, confirmed the existence of the previously theorized fourth fundamental circuit element of electrical engineering.
The existence of the new circuit, called the “memristor” (a combination of memory and resistor), was theorized back in 1971 by UC Berkeley professor Leon Chua.
The device could eventually make dynamic random access memory (DRAM) obsolete. In current systems, active computers store data in DRAM, but must shuffle the information to and from a magnetic hard disk or a flash drive, nonvolatile forms of memory. Furthermore, when the computer is turned on, the DRAM must be initially loaded from the magnetic memory. These processes consume both time and energy, slowing computing and raising the energy and heat envelopes of systems.
A memristor would need no boot up as its data would be exactly how it was previously left. Data could theoretically be read and wrote directly to and from memristors, eliminating the need for hard drives, except possibly for backup storage.
Data would be impervious to power interruptions and require a lot less power to maintain. The whole article is fascinating…I recommend it.
—–
John West is part of the team that summarizes the headlines in HPC news every day at insideHPC.com. You can contact him at [email protected].