Select News

The news in this category has been selected by us because we thought it would be interestingto hard core cluster geeks. Of course, you don't have to be a cluster geek to read the news stories.

It seems to me that we should talk about SC

ImageIf you attend SC11 you will have the opportunity to match wits in a game of Jeopardy!® with IBM's Watson computing system. Watson will be featured in a kiosk located on the 6th floor of the convention and trade center, separate from the main IBM booth. The kiosk is a version of the full IBM Power Systems-based Watson system that competed on the game show. Background on Watson can be found in our coverage of the three day Jeopardy match (day 1, day 2, day 3).

In other interesting news, the Texas Advanced Computing Center (TACC) at The University of Texas at Austin today announced that it will deploy and support a world-class supercomputer with comprehensive computing and visualization capabilities for the open science community, which seems standard enough, however, they will be using the Intel Intel® Many Integrated Core (MIC) processor to help achieve 10 petaflops of performance. The new system, called Stampede, will be built by TACC in partnership with Dell and Intel.

From the "fast compilers make me happy" department

PathScale announced today that the EKOPath 4 Compiler Suite is now available as an open source project and free download for Linux, FreeBSD and Solaris. This release includes documentation and the complete development stack, including compiler, debugger, assembler, runtimes and standard libraries. EKOPath is the product of years of ongoing development, representing one of the industries highest performance Intel 64 and AMD C, C++ and Fortran compilers. There do not seem to be any sources available just yet.

(Updated) For those that were wondering, EKOPath is a fully independent version and product. It shares a very distant past with the SGI Pro64 compiler and has evolved into a professional high performance compiler suite.

Living the terabyte lifestyle

Recently Henry Newman and Jeff Layton (distinguished ClusterMonkey) had dinner at a Cuban restaurant. While dinning, they discussed the disparity between Linux file system sizes and the growth rate of hard drives. They note:

"With 3TB, drives ext 3/4 maxes out at five disk drives. Jeff and I thought that was just insane, given you can buy five 3TB drives at Fry's and put them in your desktop. XFS maxes out at 33.3TB disk drives, and even that is far too small in our opinion. Clearly, supported file system sizes have not scaled with disk drives sizes or the demand for big data."

The full discussion, The State of File Systems Technology, Problem Statement, is at the Enterprise Storage Forum. It seems to be one of those, "wait a minute, wow, I never looked at that" moments. This issue is not just a concern for HPC or the high end issue, it really may hit home in some smaller installations. Note that the issue is not just file system sizes, but is also about performance. Henry and Jeff have a plan to explore this issue (see the end of the article).

While we are at it, read Joe Landman's What are xfs’s real limits? blog. And, check out the ensuing discussion. I respect Joe's opinion. Like me, he likes to make arguments based on hard data and benchmarks. (Like there is any other way...) We'll touch base on this again as the project continues.

Build your very own Watson Jr!

Image Just got word about a recipe for building your own IBM Watson. Of course you don't have to buy a rack of servers because Tony Pearson has outlined how to build a Watson Jr. What he suggests is small person cluster. Maybe one like this! Of course, you will loose some functionality needed to be on Jeopardy and some speed, but you can have a new friend in your basement (or garage, or bedroom). More to come. See our other Watson coverage below.

The shameless self promotion is at the end

The following news items have hit my radar screen recently. I'm sure I missed a lot of other important news, but here are some stories I found interesting. As always, there are plenty of links from which to explore further.

Speaking of HPC news, Jeff Squyres, fearless Open MPI Maven, has posted an interesting blog post called Unexpected Linux memory migration. Might want to read this one.


Login And Newsletter

Create an account to access exclusive content, comment on articles, and receive our newsletters.


Creative Commons License
©2005-2019 Copyright Seagrove LLC, Some rights reserved. Except where otherwise noted, this site is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 2.5 License. The Cluster Monkey Logo and Monkey Character are Trademarks of Seagrove LLC.