June 12, 2008
I'll be heading out to Dresden, Germany, in a couple of days to attend the annual International Supercomputing Conference (ISC) and immerse myself in all things petascale. The conference takes place from June 17-20 and HPCwire will be providing special event coverage all week ... or until I've had too much German beer.
It's my first time at the event, so I'm looking forward to seeing how it compares with the much larger U.S.-based Supercomputing Conference (SC) in November. ISC looks to be a much more academically-oriented affair than SC. It also appears to be much more manageable in size than its American version, although I'm told the conference has nearly tripled in size over the past few years.
Since ISC occurs in the middle of year, the event has become a convenient time for HPC vendors to introduce and demo new products. On the hardware side, it looks like we'll see a number of new accelerator offerings announced at ISC. In fact, with this year's introduction of the new double precision-enhanced Cell processor (PowerXCell 8i) and its use in the just-announced Roadrunner petaflop super, this might turn out to be a breakout year for HPC accelerators and heterogenous computers.
On Tuesday at ISC (technically, a pre-conference day), there are a number of presentations that I'm planning to check out. Running most of the afternoon is a series of talks on HPC requirements in the auto industry. Unfortunately at the same time is another series of talks on cluster computing that look just as interesting. I may hop back and forth between the two and cherry-pick individual presentations. I want to be sure to catch Burton Smith's talk "The Killer Micros II: The Software Strikes Back" in the cluster series.
The big opening day event on Wednesday, of course, is the announcement of the June TOP500 list. Since we're all pretty sure Roadrunner is going to be number one, that just leaves the other 499 to wonder about. The half petaflop TACC "Ranger" machine should certainly be in the top five and the recently upgraded 263 teraflop Jaguar machine at ORNL has a chance to move up from its current spot at number seven. For a more thorough breakdown of some of the other TOP500 contenders, read John Leidel's take on the list at insideHPC that he wrote back in May.
Outside of the TOP500 beauty contest, there's quite a few sessions and panels I'm looking forward to. On Wednesday afternoon, Thomas Sterling will give his annual "Year in Perspective" review of HPC. Sterling's presentation has become sort of an ISC ritual and acts as a mid-year sanity check of where the industry stands today.
On Thursday, Intel CTO Justin Ratner will give a keynote on multicore/manycore platforms for supercomputing. Probably the same stuff we've heard before, but he might give us a few extra nuggets of information about Intel's future plans. Later that day, there's going to be a two-part "Hot Seat" session for some of the big HPC vendors who agreed to subject themselves to some tough questions in exchange for a little free air time. The ISC literature describes it thusly:
In just one afternoon, 18 leading vendor representatives have ten minutes each to present their latest developments, new products and strategies. They then have to answer two tricky questions from a panel of hand-picked so-called “inquisitors”. You just have to be there to see how vendors send their best people – CTOs and even CEOs – into the arena to fight for their products.
Sounds like the makings of a reality TV show: Who Wants to Be an HPC Vendor?
On Friday, there are a couple of panels I'd like to see: the first panel is going to talk about power and cooling issues that confront the industry in the petascale era; the second panel is going to discuss the future of HPC as we head toward exascale supercomputing. A good way to end the conference before I head back home.
Posted by Michael Feldman - June 11, 2008 @ 9:00 PM, Pacific Daylight Time
Michael Feldman is the editor of HPCwire.
No Recent Blog Comments
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?