insideHPC » SC09

NASA’s Pleiades Supercomputer Among World’s Fastest

Ralph — Thu, 23 Jun 2011 18:57:57 +0000

Pleiades, NASA’s largest supercomputer, is now seventh on the TOP500 list of the world’s most powerful, high-performance computers. The announcement was made at the 26th International Supercomputing Conference in Hamburg, Germany.

Located at NASA’s Ames Research Center in Moffett Field, Calif., Pleiads supports more than 1,000 active users around the country who are advancing our knowledge about the Earth, the solar system and the universe. Pleiades is used to meet the computing needs on NASA’s most demanding modeling and simulation projects in aeronautics, earth and space science, exploration systems and technologies and future space operations.

Read the Full Story

Student cluster challenge win in case study

John West — Thu, 11 Feb 2010 13:23:46 +0000

A while ago my friend Brian from Mellanox sent me a link to a case study (PDF) they had recently put on their website. I don’t usually point to case studies, but this one is interesting in that it talks about how Mellanox worked with the team from Stony Brook University that won the Student Cluster Competition at SC09.

Each year, SC features a Student Cluster Competition where undergraduate student teams compete against each other to perform scientific application benchmark runs on clusters. StonyBrook’s student’s ingenuity won the day, aided by Mellanox 40Gb/s InfiniBand connecting Dell systems and AMD processors.

From the study

The partnership with the sponsoring vendors enables the students to develop relationships with the vendors and to learn from their engineers. The Stony Brook students built their cluster using technology provided by Dell, Mellanox and AMD. The cluster consisted of five Dell PowerEdge M905 blade servers, with 20 AMD 6-core “Istanbul” Opteron processors and 320GB of memory, all interconnected with Mellanox 40Gb/s InfiniBand.

It’s an interesting, short read.

Swimming pools. Movie stars.

John West — Mon, 14 Dec 2009 19:23:00 +0000

Our pal Joe Landman over at Scalable Informatics (an advertiser on this site) was interviewed during SC09, and has posted a link to the show

TECHiNSIGHT magazine has posted the interview. This is what they say about Joe:
“Joe Landman, founder and CEO at Scalable Informatics is a smart guy and not just because he has a PH.D. He’s developed a couple of impressive machines and come up with a great way to prove how good they are.”
In this interview, Joe discusses the I/O performance of the Scalable Informatics JackRabbit server and Pegasus workstation. He describes how we can take the data from the disks of a JackRabbit 4U server and provide it across the network at ~1.8 GB/s sustained from the disks to the application and memory, and how with multiple JackRabbits clustered together, we can multiply that 1.8 GB/s by a very large factor and deliver it to end user applications in a very trivial manner.

Cool. The video is well done, looks like they used two cameras or did some re-shooting. Also, great effects in the lower third.

Live from the show floor with the InfiniBand Trade Association

John West — Tue, 08 Dec 2009 15:23:03 +0000

During SC09 week I grabbed some time at the end of a day to sit down with Bryan Sparks and Jim Ryan in their capacities as members of the InfiniBand Trade Association (they both have day jobs; Brian is with Mellanox and Jim works for Intel). During our conversation we touched on the IBTA’s ten-year anniversary and the evolution of InfiniBand from fringe technology to stalwart of the Top500 list. Along the way we touch on some of the drivers of that adoption, the close connections of IB to PCI-e, the relationship between the The OpenFabrics Alliance and the IBTA, and how a book may hold the secret to breaking IB out of its HPC stronghold and into broader IT adoption.

Listen to the interview [audio:http://insidehpc.com/media/SC09/ibta.mp3]

Download the audio file.

Live from the show floor with Dell: a discussion about Cray, competition, and partnerships in HPC

John West — Wed, 02 Dec 2009 18:49:30 +0000

During SC week I sat down with Donnie Bell, senior manager in the enterprise marketing group, to talk about the announcement that Dell would be marketing a version of Cray’s CX1 to low-end HPC users. This audio is much more two way than some of the other conversations I’ve posted, because I was curious about some of the details that went into this deal. After giving me an overview of the system, our conversation was pretty wide-ranging over the business aspects of this deal, touching on everything from why Dell (which builds hardware) is partnering with Cray to how you get a product like this to customers who don’t identify with HPC and how the brand exposure doesn’t hurt (rather than help) Dell’s future HPC business. If you are interested in how partnerships like this happen in HPC, this is the conversation for you.

Listen to the interview [audio:http://insidehpc.com/media/SC09/dell.mp3]

Download the audio file.

Live from the show floor with Adaptive Computing

John West — Tue, 01 Dec 2009 21:10:34 +0000

As part of my audio catch-up from SC09 here is my conversation with Michael Jackson, the Co-founder, President, and COO of Adaptive Computing (formerly known as Cluster Resources). Michael is talking about Adaptive’s news during the week of the show: Moab’s connection with Voltaire’s Unified Fabric Manager and HP’s plans to resell Moab Adaptive Computing Suite, and Moab in the University of Toronto’s SciNet Consortium (interestingly, they paid for Moab in the first month based on energy savings alone).

This one has some blackberry noise for a few seconds about two-thirds of the way through — sorry. Didn’t hear it at the time.

Listen to the interview [audio:http://insidehpc.com/media/SC09/adaptive.mp3]

Download the audio file.

Live from the show floor with Avere Systems

John West — Tue, 01 Dec 2009 20:17:48 +0000

This year, as with last year, I recorded a bunch of audio during my meetings at SC. Unlike last year, however, I didn’t get hardly anything up during the show. So over the next several days I’ll be mending that sin as I work through my audio backlog.

In this segment Ron Bianchini, the President and CEO of Avere, starts off by introducing us to his well-seasoned team, and then he walks me through the story of his company and where his product is positioned in the storage acceleration market space. Avere’s appliance sits in between your storage and your server and, the company hopes, enables you to separate decisions about performance from decisions about capacity. In terms of results, here’s one: Ron walks us through Avere’s NAS storage appliance and shows results at 130,000 IOPS on the SPECsfs2008 NFS benchmark, with one quarter of the disks needed by competitors. Cool stuff.

In this audio Ron is walking me through a presentation, but the conversation is very followable without it. We recorded this in the conference registration area early Wednesday morning before the conference opened, but you’ll still hear the noise of early conference goers in the background. Hey — it’s like being there without paying extra to check your bags.

Listen to the interview [audio:http://insidehpc.com/media/SC09/avere.mp3]

Download the audio file.

Grossman on ten years of bandwidth challenges

John West — Tue, 01 Dec 2009 15:52:01 +0000

Bob Grossman, whose posts we always look forward to here at the galactic regional headquarters of insideHPC, has a new one up this AM looking back on the past ten years of the Bandwidth Challenge at SC and commenting on the increasing need for pipes to move all the data around

Some of the history is available at the web site scinet.supercomputing.org. For example, in 2000, there were 2 OC-48 (2.5 Gbps) circuits that connected the research exhibits at the conference to external research networks and the challenge was to develop network protocols and applications that could fill these circuits. The winner of the BWC (called the Network Bandwidth Challenge in 2000) was a scientific visualization application called Visapult that reached 1.48 Gbps and transferred 262 GB in 1 hour (providing 582 Mbps of sustained bandwidth utilization).

This year, there were probably at least 20 10 GE circuits that connected research exhibits to external exhibits and one of the applications reached a bandwidth utilization of over 114 Gbps.

Unofficial Stats on SC09

John West — Mon, 30 Nov 2009 19:29:57 +0000

Betsy Riley, and SC09 committee member, posted some unofficial stats on attendance at SC09 this year in response to a LinkedIn question. I’ve gotten a few emails about this, so I’m re-posting her answer here

all registrations: 10,189
tech program: 4,032
exhibitor: 3,491
exhibits only: 533 (plus 1,723 guests)
71 countries represented, all 50 states plus Puerto Rico and DC

318 exhibiting companies using 131,650 sq. feet of exhibit space
123 research exhibitors (new all-time record)
195 Industry exhibitors
largest booth ever: Microsoft, 4,500 sq. ft.

tech papers: 261 submissions, 1100 reviews, 59 accepted (22.6%)
tutorials: 28 accepted out of 71 submissions
29 panel submissions (more than twice that of SC08)

These are still only semi-official. They still have to validate the numbers (checking the list to remove duplicate registrations).
We had 95+ press badges picked up–an all time high.

SC09 flash storage panel video

John West — Fri, 27 Nov 2009 16:09:03 +0000

Rich Brueckner has posted a link to video they shot during SC09 at the flash storage panel

Abstract: With an exponential growth spurt of peak GFLOPs available to HPC system designers and users imminent, the CPU performance I/O gap will reach increasingly gaping proportions. To bridge this gap, Flash is suddenly being deployed in HPC as a revolutionary technology that delivers faster time to solution for HPC applications at significantly lower costs and lower power consumption than traditional disk based approaches. This panel, consisting of experts representing all points of the Flash technology spectrum, will examine how Flash can be deployed and the effect it will have on HPC workloads.

Video here.

Birds do it, bees do it

John West — Fri, 27 Nov 2009 04:27:16 +0000

Well everyone else is putting pen to paper on what they did with their time at SC, so here’s mine. I’ll skip the things that everyone else is already writing about, and jump to the few observations I have that I haven’t seen elsewhere.

During the course of the conference I had 45 meetings with the companies, programs, individuals, and agencies who are shaping the direction our community is headed in both the near and far term. In the areas of the conference I was trolling, the emphasis was decidedly short term: the 2020 exascale challenge was an undercurrent, but not an overarching theme. Dan Reed had a different experience, probably because he actually spent time in the tech program (a scheduling choice I intend not to repeat next year). On my promenades around the floor, I saw two trends.

Into pieces and parts

First, there were several companies presenting technical solutions that allow datacenter managers to break up the traditional concept of a compute node, freeing them to compose systems from pieces and parts at a finer level of granularity. Companies such as RNA networks, 3Leaf Systems, Avere Systems, and NextIO all occupy different segments of this trend. NextIO is an exemplar. Their solution allows cluster builders to deploy compute nodes with only local memory and a processor, leaving the IO — from network ports to GPUs and drives — to be deployed top-of-rack in a separate chassis that is PCI Express connected to nodes that need access to those resources. As the company says, this lets you separate decisions about IO from decisions about compute.

A use case? Rather than adding GPUs to just a few servers in your small cluster you could add them in a NextIO device that would make them available to all the servers. The point is not that all the servers would use them simultaneously (that would be bad), but rather that you wouldn’t have to decide a priori which servers would get the GPUS, allowing you to avoid idle resources held “just in case” a GPU user showed up.

The other providers in this market offer solutions that focus on memory virtualization, or combinations of memory and disk. As a whole they are beginning to gain traction with the major vendors, and most have announced partnerships with at least some of the Tier 1 OEMs.

From GPUs to accelerators

The other major technical theme I saw was the beginning generalization of the GPU. Two years ago custom accelerator company ClearSpeed was the darling of SC — it had disappeared from the conversation less than a year later, supplanted by the juggernaut that is GPGPU-based technical computing. GPU-manufacturer NVIDIA has incredible momentum right now, but its useful to remember that what goes up also may go down. An early reminder of this is that a significant fraction of the attendees and exhibitors I spoke with at SC09 are already thinking of GPGPUs as a specific case of accelerated computing, and aren’t referring to GPUs at all. NVIDIA is facing a strong first-mover disadvantage, and there is a real risk that others will come into the market with a different technology implemented in a better way. Also, there is a growing number of developer tools that support expression of parallel work that can be mapped to a variety of accelerators, including but not limited to GPUs (The Portland Group is an example, as is OpenCL). As developers adopt these tools their mobility to new technologies will be facilitated, to NVIDIA’s potential disadvantage.

NVIDIA is also facing business risk: with Intel moving to bring high performance graphics capabilities directly into the CPU, NVIDIA may find it’s volume graphics board business under siege. The primary attribute that makes NVIDIA’s GPUs attractive to HPC users is their very low price/performance. If NIVIDA loses the commodity multiplier that allows them to keep the price so low on its GPUs, the HPC community will look for other alternatives. In HPC, “good enough” is good enough, and volume always wins.

SC09 reflections from Dan Reed

John West — Fri, 27 Nov 2009 03:50:09 +0000

In addition to SC09 reflections from our own John Leidel, Dan Reed has posted his own reflections on the conference. As always, his thoughts are worth a read. He does make a comment that I found surprising

As always, a new Top500 list was revealed, and exascale computing was (literally) a hot topic, as multiple groups discussed the energy and reliability challenges inherent in building exascale systems. Jack Dongarra and collaborators summarized the results of the International Exascale Software Project (IESP), including a putative roadmap produced following the most recent workshop. A group of national laboratories and universities also announced the Hybrid Multicore Consortium (HMC), whose goal is to “address the migration of existing applications to accelerator-based systems and thereby maximize the investments in these systems.”

I commented in my own writeup — which I’ll post soon — that I actually hadn’t heard much talk about exascale. This is probably solely due to the different paths Dr. Reed and I took during the conference. I spent all my time on the exhibits floor and meeting with HPC ecosystem companies, and I suspect he spent a good bit of time in the tech program sessions. I believe that next year I’ll reverse my outward orbit from the tech program and devote my time attending (and live blogging) the tech sessions.

Also, Dan gives us a peek behind the curtain at the conference

Dona Crawford (LLNL), Wilf Pinfold (Intel and SC09 conference chair) and I recruited the panel of plenary and keynote speakers: Justin Rattner (Intel CTO), Lee Hood (Institute for Systems Biology) and former U.S. Vice President Al Gore, who gave the conference keynote. (Dona deserves full credit for landing al Gore, as she worked tirelessly to negotiate terms, processes and topics.)

I’ll say again that I think the choice of Al Gore was brilliant. I guess I should have shouldered my way up to Dona too when I thanked Wilf. Sorry for the ommision, Dona.

SC09 Debrief: software is king but people rule all

John Leidel — Wed, 25 Nov 2009 17:06:07 +0000

It’s become an annual tradition for John Leidel (aka, “the other John”) to write up his thoughts following SC. Submitted for your approval, the product of this year’s ritual.

Following my annual pilgrimage to the wondrous land of supercomputing tchotchkes, I spend a few days reflecting on what I saw, what I heard and most importantly, what I learned. Regardless of how hard vendor marketing departments work to trump one another, there always seems to exist a resonating buzz around a small subset of technologies, events or HPC ideologies. This year was no exception.

Software, software, software

SC09 saw its fair share of major product announcements. There was the first public sighting of SGI’s Ultraviolet, Cray released details on the next iteration of the XT series, Appro remained extreme with a new Xtreme-X platform, and NVIDIA opened the kimono on Fermi. However, it was clear by Tuesday afternoon that the hardware news was being overshadowed by the various software technologies on display.

After taking a walk through the poster sessions, this was clearly the year of accelerated computing software stacks. The NVIDIA CUDA following was well-represented in poster sessions, BOFs and paper presentations. I’m not sure of the exact distribution of submissions, but it certainly seemed as though CUDA dominated the technical submissions more than any other single technology. Allinea and TotalView Tech announced debugging utilities for CUDA and EM Photonics spelled it out with LAPACK support.

GPUs not strike your fancy? There was plenty of news for the FPGA crowd as well. The industry buzz was definitely thick around the news from Convey. The hybrid computing startup surprised many of us with news of progress on both the business and technology fronts: Convey announced two dozen ships (including some very high profile customers), and showed off their financial analytics FPGA personality. The Toal-Bewer-Wallach trifecta is HPC business savvy and it shows. Mitrionics was also on hand to surprise quite a few folks with their latest compiler R&D.

As the week moved on it also became clear that people are starting to worry seriously about the various software implications with building accelerator-based applications at scale. Accelerated computing technologies are beginning to move closer to the center and traditional scalar technology is beginning to become more extreme. We, as an industry, need to begin developing more cohesive strategies sufficient to drive the industry to its next major milestone: exascale computing.

People: The Cornerstone to Success

While squeezed into coach on the flight from home to Portland, I had an interesting thought: will the global financial strife affect the attendance and quality of SC? Quite the contrary, we had a great number of people attend the show from every niche of HPC. In addition to the good showing, an interesting phenomenon swept through the side halls and party conversations. “We’re Hiring.”

Through friendly conversations with the gray beards haunting the most powerful reaches of HPC, I found that many of the most coveted organizations in our industry are in desperate need of supercomputing wizards to work their black magic on a growing number of super-scale systems. NOAA, DoE, DoD and NSF-sponsored labs alike are looking to pick up skilled individuals to operate their big ticket machines. So why not just hire the right people? The key phrase here is skilled.

According to those in need, the pool of truly skilled talent in our fair burg of HPC is shallow and shrinking quickly. The individuals that have been driving the innovation in this industry for the past thirty years are continuing to do so without the bench necessary to replace them at age 65. As such, it seems we’re retiring talent faster than its being replaced.

We should learn two things from our current situation. First, people are the driving force behind success. Silicon procurements, interconnects and compilers are nothing without the code ninjas and keyboard cowgirls necessary to keep the gears turning. Second, given the amount of talent concentrated in the upper echelons of our industry, its vital for the true technologists among us to develop their mentor skills and pass on their knowledge to those below.

Until Next Year

All in all, I think SC09 was a success. Despite the global financial crisis, we had thousands of folks brave the cold winds of Portland to attend. As always, good food and good friends were the highlight of my trip. I hope to see everyone next year in New Orleans for SC10.

StarGate Demo at SC09 Shows How to Keep Astrophysics Data Out of Archival “Black Holes” [UPDATED with pics]

Jon Bashor — Fri, 20 Nov 2009 23:22:35 +0000

Reserved Bandwidth on ESnet Makes Possible Mulit-Gigabit Streaming Between Argonne, SC Conference in Portland

[UPDATE: I published this on the plane, and didn't have enough bandwidth to add the pics. They're inline now.]

As both an astrophysicist and director of the San Diego Supercomputer Center (SDSC), Mike Norman understands two common perspectives on archiving massive scientific datasets. During a live demonstration at the SC09 conference of streaming data simulating cosmic structures of the early universe, Norman said that some center directors view their data archives as “black holes,” where a wealth of data accumulates and needs to be protected.

But as a leading expert in the field of astrophysics, he sees data as intellectual property that belongs to the researcher and his or her home institution — not the center where the data was computed. Some people, Norman says, claim that it’s impossible to move those terabytes of data between computing centers and where the researcher sits. But in a live demo in which data was streamed over a reserved 10-gigabits-per-second provided by the Department of Energy’s ESnet (Energy Sciences Network), Norman and his graduate assistant Rick Wagner showed it can be done.

While the scientific results of the project are important, the success in building reliable high-bandwidth connections linking key research facilities and institutions addresses a problem facing many science communities.

“A lot of researchers stand to benefit from this successful demonstration,” said Eli Dart, an ESnet engineer who helped the team achieve the necessary network performance. “While the science itself is very important in its own right, the ability to link multiple institutions in this way really paves the way for other scientists to use these tools more easily in the future.”

“This couldn’t have been done without ESnet,” Wagner said. Two aspects of the network came into play. First, ESnet operates the circuit-oriented Science Data Network, which provides dedicated bandwidth for moving large datasets. However, with numerous projects filling the network much of the time for other demos and competitions at SC09, Norman and Wagner took advantage of OSCARS, ESnet’s On-Demand Secure Circuit and Advance Reservation System.

“We gave them the bandwidth they needed, when they needed it,” said ESnet engineer Evangelos Chaniotakis. The San Diego team was given two two-hour bandwidth reservations on both Tuesday, Nov. 17, and Thursday, Nov. 19. Chaniotakis set up the reservations, then the network automatically reconfigured itself once the window closed.

At the SDSC booth, the live streaming of the data drew a standing-room-only crowd as the data was first shown as a 4,096³ cube containing 64 billion particles and cells. But Norman pointed out that the milky white cube was far too complex to absorb, then added that it was only one of numerous time-steps. In all, the data required for rendering came to about 150 terabytes of data.

In real time, the data was rendered on the Eureka Linux cluster at the Argonne Leadership Computing Facility and reduced to one-sixty-fourth of the original size for a 1,024³ representation, making it more manageable and able to be explored interactively. The milky mesh was shown to contain galaxies and clusters linked by sheets and filaments of cosmic gases. Its all clearer in the movie, which you can see here (other resolutions at the bottom of this page).

The project, Norman explained, is aimed at determining whether the signal of faint ripples in the universe known as baryon acoustic oscillations, or BAO, can actually be observed in the absorption of light by the intergalactic gas. It can, according to research led by Norman, who said they were the first to determine this. Such a finding is critical to the success of a dark energy survey known as BOSS, the Baryon Oscillation Spectroscopic Survey. The results of his proof-of-concept project, Norman said, “ensure that BOSS is not a waste of time.”

Creating a simulation of this size, even using the petaflops Cray XT5 Kraken system at the University of Tennessee can take three months to complete as it is run in batches as time is allocated, Norman said. The data could then be moved in three nights to Argonne for rendering. The images were then streamed to the SDSC OptiPortal for display.. Norman said the next step is to close the loop between the client side and the server side to allow interactive use. But the hard work — connecting the resources with adequate bandwidth — has been done, as evidenced by the demo, he noted.

But it wasn’t just an issue of bandwidth, according to ESnet’s Dart. “We did a lot of testing and tuning,” said Dart. ESnet is managed by Lawrence Berkeley National Laboratory (LBNL).

Other contributors to the demo were Joe Insley of Argonne National Laboratory (ANL), who generated the images from the data, and Eric Olson, also of Argonne, who was responsible for the composition and imaging software. Network engineers Linda Winkler and Loren Wilson of ANL and Thomas Hutton of SDSC worked to set up and tune the network and servers before moving the demonstration to SC09. The project was a collaboration between ANL, CalIT2, ESnet/LBNL, the National Institute for Computational Science, Oak Ridge National Laboratory and SDSC.

CULA 1.1 released, SC09 giveaway for insideHPC readers success

John West — Fri, 20 Nov 2009 23:09:28 +0000

Liana at EM Photonics emailed me to let me know that the promo they did for insideHPC readers exceeded our expectations. By the time that day 2 started the fifth copy was in the hands of some eager code monkey hoping to accelerate his or her app with CULA’s GPU-accelerated LAPACK routines.

She also let me know they’ve revved CULA to version 1.1. There isn’t a press release, but this is what she sent via email

We’re very excited to announce the release of CULA 1.1 Beta. This new version of CULA includes improvements from every angle. Here is a subset of the improvements that have made it into this release:

Exciting new functions including general Eigensolver (CULA Premium Feature)

Bridge interface for migrating currently existing LAPACK/MKL code

Better documentation including a full API reference

New examples constructed from user feedback

More performance!

Mac OS X support (Preview)

In this new version of CULA, we’ve put a lot of effort into responding to users questions, comments, and suggestions. One of the most common suggestions has been to provide eigenvalue calculations, as that was one of the major subject areas of linear algebra our 1.0 release hadn’t provided – and so we ensured in this release that this functionality was our top priority.

Another common suggestion has been to provide a migration path from other LAPACK systems to CULA. With CULA 1.1, we’re introducing one of our most exciting new features: the Bridge interface. This interface is targeted at users who are porting existing linear algebra codes to a GPU-accelerated environment. It makes this job easier by matching the function names and signatures of several popular linear algebra packages including MKL, ACML, and cLAPACK. It additionally provides a fallback to CPU execution when a user does not have a GPU or when problem size is too small to take advantage of GPU execution. By providing a migration path from three major LAPACK packages, we hope this feature shows how committed we are to making CULA easy to use.

One of the things that users have asked for is a detailed function reference. Although CULA closely mirrors other LAPACK packages, some of our users (especially those without LAPACK experience) were confused by some functions and semantics. With this release, we’ve included a reference manual that is a companion to the programmers guide. We’ve put a lot of effort into making sure that these documents cover all of the questions our users have posted and we hope it shows in this new release.

With CULA 1.1, we’ve created 3 new examples. One of these examples showcases the Bridge interface mentioned I mentioned previously. The other example demonstrates using CULA in a multi-threaded multi-GPU environment, a common question our users have asked on our forums. The last example uses system solves to demonstrate using CULA with data types other than single-precision real. If you have any suggestions for new examples, post your suggestions in the forums and we may be able to get these into the final release of 1.1.

Last but not least, we’re introducing a Preview build for Mac OS X 10.5 Leopard. Why do we say it’s a Preview? Currently, our Mac toolchain is limiting us to supporting only single-precision real across the board, but we’ll be expanding this support very shortly. Please keep an eye on this blog for future updates; we have a lot to say about Mac!

NextIO announces partnership to get IO virtualization tech into IBM clusters

John West — Fri, 20 Nov 2009 22:41:39 +0000

NextIO is a company with a cool offering that allows you to pool your network ports, disk drives, GPUs, and other IO devices into a box that can be connected via PCI Express to multiple servers. As the company says, this lets you separate decisions about IO from decisions about compute. A use case? Rather than adding GPUs to just a few servers in your small cluster you could add them in a NextIO device that would make them available to all the servers. The point is not that all the servers would use them simultaneously (that would be bad), but rather that you wouldn’t have to decide a priori which servers would get the GPUS, allowing you to avoid idle resources held “just in case” a GPU user showed up.

This week they announced a partnership with IBM that puts their gear into IBM clusters for users that can take advantage of it

NextIO, a premier provider of next-generation I/O solutions, today announced it is working with IBM to offer customers integrated cluster solutions that incorporate NextIO technology, with availability in 2010. The solution will enable reconfigurable and on demand GPU compute capabilities for IBM iDataplex customers. The announcement was made at the Supercomputing 2009 show in Portland, Ore.

…The GPU virtualization solution will offer the ability for a single IBM iDataPlex™ server to access from one to eight double-wide GPUs or up to 16 single-wide GPUs in the appliance. Users can quickly and easily enable more or less GPU resources on demand, depending on application requirements. Each iDataPlex rack supports 10 GPU appliances providing up to 160 GPUs and over 80TFlops of compute processing per rack.

Pics added to community leadership award presentation story

John West — Fri, 20 Nov 2009 22:33:33 +0000

Just a quick note: I’ve added a couple pics we had shot during the presentation of the inaugural HPC Community Leadership Awards to Bill Gropp and ORNL to my original story. I tried uploading larger versions from the plane, but the connection isn’t reliable enough. Will upload those on the ground.

Awards at SC09

John West — Fri, 20 Nov 2009 22:25:40 +0000

In case you didn’t notice all of the little foam core signs hanging in booths all over the show floor, HPCwire announced its raft of reader and editor awards again during SC09. They announced winners in ten categories. More in their article here. Congratulations to all the winners.

And, of course, we announced that Bill Gropp and Oak Ridge were voted by the community (1,000 votes in each category!) to receive the inaugural HPC Community Leadership Awards.

Was Verari at SC09?

John West — Fri, 20 Nov 2009 22:17:20 +0000

Anyone know if Verari was on the show floor at SC09? I must have missed them…they were scheduled to be there, but I didn’t see the booth.

A look at the Green500: IBM, and a couple other vendors

John West — Fri, 20 Nov 2009 22:11:40 +0000

First off, if you haven’t been by the Green500 website in a while, swing by. They’ve gotten a full facelift, and they look great. Now let’s take a quick look at the latest lists.

On the TOP Green500 list, which ranks the Top500 entries in order of power efficiency (MFLOPS/W) not Linpack performance, we see IBM leading the list again this year with its PowerXCell 8i offering (I think this is the QS22 blade — comments anyone?). That system takes you through the top 5 slots. Slots 6 and 7 are occupied by custom systems (GRAPE-DR and the Xeon/ATI cluster), and then a run of 15 systems at various positions all owned by IBM’s BlueGene/P. The run is interrupted by a system from Sun and NEC, and then there is a boatload of iDataPlex’s.

The Little Green500 broadens the definition of a supercomputer to help guide purchasing decisions for smaller institutions. According to the website, to be eligible for this list you’ve got to have a system as fast as the 500th ranked supercomputer on the TOP500 list 18-months prior to the release of the Little Green500.

Who leads this list? IBM baby, with QS22′s and BlueGene/P’s, and iDataPlex’s again. IBM doesn’t just lead in energy efficiency — they own it.