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
– SGI and Mitrionics recognized for their-FPGA accelerated BLAST calculations, http://www.sgi.com/company_info/newsroom/press_releases/2007/march/bioteam.html
– Computerworld article profiles benefits of HPC to old-world industry, http://www.computerworld.com/action/article.do?command=viewArticleBasic&articleId=9012680
– NC State professor builds cluster from 8 PS3s running Linux, http://gaming.engadget.com/2007/03/10/nc-state-engineer-crafts-academic-cluster-with-playstation-3s
– IBM announced that they've started making Cell processors at 65nm,
http://insidehpc.com/2007/03/13/cell-at-65nm/
– Buy the bubble: bits of the Intel Shell for $20, http://insidehpc.com/2007/03/15/buy-a-piece-of-the-bubble-the-intel-shell-for-sale/
>>HPC R&D Act passes House
The HPC R&D Act, which we've talked about before, passed the house on a voice vote yesterday according to the CRA's Policy Blog:
“The bill would amend the original High Performance Computing and Communications Act of 1991 (HPCC) to attempt to provide sustained, transparent access for the research community to federal HPC assets, assure a balanced research portfolio and beef up interagency planning.”
Previous versions of this bill died in the Senate during the 108th and 109th Congresses. After talking with Senate staffers the CRA feels this might be the year. Find a PDF summarizing the impact on current law at http://www.cra.org/govaffairs/blog/archives/HPCC_Act_as_amended_byHR1068.pdf.
>>Virtualization and hosted hardware/software solutions continue to call attention to themselves
Sun has been recently talking more about hosted computing solutions, including hosted applications for science and engineering that ordinarily run locally on supercomputers. All of this ties in with their grid computing and virtualization strategies. Scott McNealy talks about the concept in this PDF: http://tinyurl.com/2y9rkt.
Sun announced this week that it has added additional features to its Network.com pay-per-use offering, giving developers and open source communities the ability to just “click and run” applications online and a unique way to create personalized application playlists. ISVs develop and publish applications online, and customers pay per use, running the apps on Sun's big back end machines. No software to buy, no hardware to buy.
The offering includes some significant science and engineering apps and early customers listed by Sun include some traditional HPC users like Lawrence Livermore and the EPA who are either kicking the tires or satisfying some of their surge demand.
HPC and supercomputing aren't words we typically see in the broader IT press, but even IT World Canada has picked up on all the virtualization activity right now and identified 2007 as the year that supercomputing gets to the “masses” in their Top 5 IT Trends for 2007 (http://tinyurl.com/ywt3h6). I think we could be on this path, but I predict 2007 is not going to be a “breakout” year. HPC is still too hard to {program, execute, understand} for the “masses.”
>>Europe focuses on speed
Primeur Weekly is running an article on the report of the HPC in Europe Taskforce (HET), an initiative of eleven European countries. There looks to be some sound reasoning here, for example the desire to foster a robust European HPC ecosystem.
There's also a focus on software, which is what all the kids on this side of the pond are talking about these days. But then there's this:
“European computational science needs computing capacities that match the fastest systems in the USA and Japan, HET says. It is an intolerable situation for the competitiveness of European computational science that this currently is not the case.”
At least the press around this looks like another HPC arms race. In fact, the HET's number 1 recommendation is to procure “extreme computing power.” The bottom line is that the US and Japan are “solve it with money” cultures, so Europe probably won't ever win in a FLOPS arms race.
But why try? Take the lead on getting actual work done and beat the pants off the HPC superpowers with a resurgence in your own national competitiveness.
>>Wing: software in the US in big trouble
Remarks by Jeannette Wing, head of the CS department at Carnegie Mellon, posted last week at Computerworld.com:
“Wing sees her job as making esoteric research issues real, immediate and relevant to users. 'Today in security, we are patching systems and fighting viruses and worms and doing source code analysis using techniques that the basic research community invented 20 years ago, or even longer than that,' she said.
“Consequently, users 'are basically putting Band-Aids on our software and trying to build better intrusion-detection systems,' said Wing, while the basic research needed to protect against future threats isn't being done.”
Dr. Wing is set to take over NSF's computing research spending on July 1. The NSF funds 87 percent of all federally funded fundamental CS research out of a budget that runs a paltry $530M.
—–
John West summarizes the headlines in HPC every day at insideHPC.com, and writes on leadership and career issues for technology professionals at InfoWorld and on his own blog at onlytraitofaleader.com. You can contact him at [email protected].