Visit additional Tabor Communication Publications
September 22, 2011
The HPC contingent at NVIDIA must be stinging a bit today when they learned of Intel's first big Many Integrated Core (MIC) win at the Texas Advanced Computing Center (TACC). The 10-petaflop super, named Stampede, is scheduled to boot up at the end of 2012, and will have 8 petaflops worth of Intel's manycore Knights Corner coprocessors.
The NSF, who funded the system, is shelling out $2.5 million for thousands of MIC chips -- money that would probably be going to NVIDIA's bottom line had Intel stayed out of the HPC coprocessor biz. More disturbing for NVIDIA perhaps is that Intel doesn't even have production parts in the field, but managed to convince the NSF, TACC and its partners that MIC was the way to go. When I spoke with TACC Director Jay Boisseau, he told me that the manycore chips they expect to be getting in the fall of 2012 are pre-production Knights Corner parts, but otherwise identical to the final version.
TACC telegraphed the deal in April when it signed up to port some science apps to Knights Ferry, the 32-core MIC prototype platform. They subsequently built a small cluster outfitted with Knights Ferry coprocessors and started kicking the tires. Apparently that went well enough to warrant the 8 petaflop buy-in.
NVIDIA managed to garner some token approval in the Stampede project. A hundred or so of its next-generation Kepler GPUs will be hooked into some of the nodes, but mostly to be used for remote visualization. There are also a couple of NVIDIA GPUs attached to each of the 16 nodes of Stampede's big memory sub-cluster for data analytics-type workloads.
But the lion's share of the supercomputer's computational horsepower will come from Intel silicon. In addition to the 8 petaflops of MIC, 2 petaflops will be supplied by the upcoming Sandy Bridge CPUs, the 8-core Xeon E5, which will power the main cluster of 6,400 dual-socket nodes. An additional 5 petaflops of second-generation MIC coprocessors will be added in a couple years.
Knights Corner is going to have upwards of 50-cores and deliver something north of 1 teraflop of double precision number crunching performance, which should more or less match the new Kepler GPUs. The new NVIDIA parts are slated to arrive sometime in 2012, and it looks like volume production for Knights Corner won't start until 2013. I don't know how much of difference that will really make, especially in light of the choice TACC and the NSF just made.
Intel has more than 100 partners right now running their codes on the prototype MIC processors. The early returns appear to be positive, especially in regard to ease of programming. The x86 manycore architecture lends itself well to OpenMP-style programming, which gives it a built-in HPC customer base.
That's something NVIDIA didn't have when it began its GPGPU push into HPC in 2006. The GPU maker had to invent the CUDA software framework and then bring developers on board, which it managed to do quite successfully over the last five years. Intel, with its OpenMP-friendly architecture, will allow many HPC developers to leapfrog much of that onerous software transition phase.
Intel is pushing hard into HPC right now, noting opportunities like Stampede to sell barrels of chips with a single deal. But there is also the general feeling that the HPC market is an insatiable beast for high-powered processors. In a blog posted on Thursday by Joe Curley, Intel Director of Technical Computing, he lays out the rationale for Intel's high performance computing enthusiasm:
Last week at the Intel Developer Forum, Kirk Skaugen, VP and GM of Intel’s Data Center and Connected Systems Group, talked about the huge growth that is expected in HPC in coming years. (For those who didn’t have a chance to attend IDF you can see video of Kirk’s presentation here -> part 1 & part 2). Our estimations show that by 2015, the world’s top 100 supercomputers will be powered by 2 million CPUs and by 2019 this number will reach 8 million CPUs. To give you a perspective, in 2010, Intel shipped about 8 million server processors in total.
That's a rather compelling incentive for any chipmaker to be in the HPC business.
Posted by Michael Feldman - September 22, 2011 @ 6:01 PM, Pacific Daylight Time
Michael Feldman is the editor of HPCwire.
No Recent Blog Comments
Contributing commentator, Andrew Jones, offers a break in the news cycle with an assessment of what the national "size matters" contest means for the U.S. and other nations...
Today at the International Supercomputing Conference in Leipzing, Germany, Jack Dongarra presented on a proposed benchmark that could carry a bit more weight than its older Linpack companion. The high performance conjugate gradient (HPCG) concept takes into account new architectures for new applications, while shedding the floating point....
Not content to let the Tianhe-2 announcement ride alone, Intel rolled out a series of announcements around its Knights Corner and Xeon Phi products--all of which are aimed at adding some options and variety for a wider base of potential users across the HPC spectrum. Today at the International Supercomputing Conference, the company's Raj....
Jun 18, 2013 |
The world's largest supercomputers, like Tianhe-2, are great at traditional, compute-intensive HPC workloads, such as simulating atomic decay or modeling tornados. But data-intensive applications--such as mining big data sets for connections--is a different sort of workload, and runs best on a different sort of computer.
Jun 18, 2013 |
Researchers are finding innovative uses for Gordon, the 285 teraflop supercomputer housed at the San Diego Supercomputer Center (SDSC) that has a unique Flash-based storage system. Since going online, researchers have put the incredibly fast I/O to use on a wide variety of workloads, ranging from chemistry to political science.
Jun 17, 2013 |
The advent of low-power mobile processors and cloud delivery models is changing the economics of computing. But just as an economy car is good at different things than a full size truck, an HPC workload still has certain computing demands that neither the fastest smartphone nor the most elastic cloud cluster can fulfill.
Jun 14, 2013 |
For all the progress we've made in IT over the last 50 years, there's one area of life that has steadfastly eluded the grasp of computers: understanding human language. Now, researchers at the Texas Advanced Computing Center (TACC) are utilizing a Hadoop cluster on its Longhorn supercomputer to move the state of the art of language processing a little bit further.
Jun 13, 2013 |
Titan, the Cray XK7 at the Oak Ridge National Lab that debuted last fall as the fastest supercomputer in the world with 17.59 petaflops of sustained computing power, will rely on its previous LINPACK test for the upcoming edition of the Top 500 list.
05/10/2013 | Cleversafe, Cray, DDN, NetApp, & Panasas | From Wall Street to Hollywood, drug discovery to homeland security, companies and organizations of all sizes and stripes are coming face to face with the challenges – and opportunities – afforded by Big Data. Before anyone can utilize these extraordinary data repositories, however, they must first harness and manage their data stores, and do so utilizing technologies that underscore affordability, security, and scalability.
04/15/2013 | Bull | “50% of HPC users say their largest jobs scale to 120 cores or less.” How about yours? Are your codes ready to take advantage of today’s and tomorrow’s ultra-parallel HPC systems? Download this White Paper by Analysts Intersect360 Research to see what Bull and Intel’s Center for Excellence in Parallel Programming can do for your codes.
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?
Join our webinar to learn how IT managers can migrate to a more resilient, flexible and scalable solution that grows with the data center. Mellanox VMS is future-proof, efficient and brings significant CAPEX and OPEX savings. The VMS is available today.