The Week in Review
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
Supermicro launches new blades, denser than HP and Sun;
InfoWorld review on Sun Fire X4500 server: 48 drives in 4U;
Grid Service Broker 3.1 released by U of Melbourne;
Voltaire offers 20 Gbps IB switch aimed at HPC;
NSF pencilled in for 10 percent increase, so far so good;
>>Evergrid launches new job management tools, partners with Platform
Evergrid, which released its Cluster Availability Management Suite (CAMS) this week, has also announced a partnership with Platform Computing to integrate its Availability Services with LSF.
Evergrid provides application fault tolerance using an OS abstraction layer that loads between the operating system (OS) and the application.
Without modifying either the application or the operating system, CAMS/AvS periodically captures the collective state of the application while it's running. By recording the state of an application and all of the OS and system state, Evergrid says that it is able to checkpoint and resume from failures or interruptions rapidly with “minimal” overhead.
This is good for long running user applications obviously, but it also provides something very powerful for HPC center operators: preemptive scheduling. Because jobs can be suspended and returned to execution on command, centers can now be a lot more creative with batch scheduling policies without sacrificing high utilization numbers.
This is great news for users who want bigger machines but are plagued by the instabilities that come with them, and great news for HPC centers that need to accommodate 2 CPU jobs alongside 10,000 CPU jobs without sacrificing utilization.
You can read more, along with links to the company's info on these announcements, at http://insidehpc.com/2007/06/13/evergrid-and-platform-partner/ and http://insidehpc.com/2007/06/12/evergrid-launches-new-cluster-availability-tool/.
>>IBM's blades for the little guy
IBM announced its new BladeCenter S this week. The new blade computing system is aimed at allowing smaller businesses to integrate the servers used in a small office environment into one system that can sit on a desktop.
According to the company the system is designed to be deployed outside of a machine room, plug into a standard 110 volt power outlet, and manage storage and up to six blade servers at a time.
The BladeCenter S is expected to be available in Q4 of 2007. Although it's targeted for the enterprise crowd, this little guy could easily be a departmental-level HPC box for small R&D environments. The iPod of IBM's HPC offerings.
And — you guessed it — there's even an energy advantage. According to IBM the BladeCenter S can reduce a company's server energy use up to 85 percent. Green really is the new black. More from IBM at http://www-03.ibm.com/press/us/en/pressrelease/21712.wss.
>>Japan's super super
Japan is targeting a 10 PFLOPS computer in 2011. The interesting bit is the team that RIKEN has put together to build it includes rivals Hitachi, NEC, and Fujitsu.
From the story at Tech-On!:
“The Working Group for the Assessment of the Next-Generation Supercomputer Framework Design under Japan's Ministry of Education, Culture, Sports, Science and Technology (MEXT) has summarized and disclosed an 'assessment report' about the architecture of the next-generation supercomputer being developed toward the goal set in FY2011.
The report predicts that the next-generation supercomputer's architecture will be a multi-system combining an accelerator-embedded scalar computer and a vector computer.”
The report also cites a specific, testosterone-laden goal: “winning the top position in June 2011” on the Top500 list.
The architecture looks a lot like what Cray and others are talking about with respect to heterogeneous systems: a scalar processor-based computing system added with an accelerator is combined with a vector processor-based computing system.
Check out the full story at
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 firstname.lastname@example.org.