HPC

Powerful ARMs, heaving racks ... yes, of course it's this year's student cluster-building contest

Meet the ISC'16 Student Cluster warriors


HPC Blog It's time for our traditional video look at each of the teams in the ISC (International Supercomputing Conference) 2016 Student Cluster Competition. Let's take it alphabetically this time...

Team Boston Green: They're one of two teams utilizing ARM processors to fuel their cluster. A lot of ARM processors, about 800 cores worth. They've also added on eight NVIDIA Jetsons to give a little numerical processing punch to their box. In the video we talk about their system and a little team bonding.

Youtube Video

Team Estonia: No team at ISC'16 had more trouble than Team Tartu. They were using an entirely new architecture (Power 8), hadn't had much time with their system, and didn't have all of the resources they needed to make it perform. Other than that, things were OK. In the video, we talk about their experiences so far and the biggest challenges they're facing in the competition.

Youtube Video

Team Hamburg: Team Hamburg is loaded for bear at ISC'16, at least when it comes to LINPACK. They're driving a very small configuration at only two nodes, but with a ton (eight) of NVIDIA K80 accelerators. This kind of configuration tends to make me think they’re going after the LINPACK award, but it could also, depending on the apps and team skills, be a strong competitor on the non-LINPACK stuff too. As you'll see on the video, Team Hamburg also has a cool approach to their rack.

Youtube Video

Team NERSC: The first all-female team to compete at an ISC competition is pushing a Cray blade cluster. In the video, we get to know the team and find out some interesting facts. First, each of them interned at NERSC at some point in time – including the two team members who are still in high school. Check out the video to learn more about them.

Youtube Video

Team Purdue/Colorado: While these two teams have more experience as institutions than any others, the students on the team are mostly new to the wild world of international clustering. In the video, I try to stir the pot and get some team dissension going, but I'm unsuccessful. So far, the team thinks that the Graph500 application (where they have to use their own algorithm) will be the most challenging.

Youtube Video

Team Singapore: Nanyang Technological University of Singapore is another team with a smallish configuration. While they only have three nodes and 72 cores, they are counting on their six NVIDIA K80s to provide some hardcore compute. The team explains their rationale for their configuration in the video, along with some chitchat about Graph500, MPI, and cache utilization.

Youtube Video

Team Shanghai: This team got their ticket punched by coming in second place at the ASC'16 cluster competition. However, they weren't happy with it – they wanted to win the damn thing. They're not letting the HPC luminaries or the heady sights and sounds that go with an ISC conference get to their heads; they're concentrating on their cluster competition game.

Youtube Video

Team South Africa: Even though this is Team South Africa's fourth appearance (taking home two championships and one second place finish) at the ISC Championship, it's all new to this team. South Africa has a policy that requires each of their international teams to be made up of entirely new members. In the video, there's a decidedly one-sided conversation in which the team captain, with the team arrayed behind him, explains their approach to the competition. We talk in some detail about their Graph500 effort, and how the team is geared to win the whole competition – rather than HPL or HPCC. As he puts it, "we're hoping to dominate on the other benchmarks."

Youtube Video

Team Spain: The kids from Spain have the fever – ARM fever. For this, their second competition, they've cranked up the ARMiness of their cluster to the point where they're sporting more than 800 cores that are liquid cooled, courtesy of Cool IT. The team admits to making a rookie mistake in accidentally launching two simultaneous HPL runs (which will certainly tamp down your performance). In the video, we also talk about how the team is tackling the other apps.

Youtube Video

Team Tsinghua: As usual, Team Tsinghua is both articulate about their strategy and confident of their chances when it comes to the competition. In the video, we talk about their choice of the consumer NVIDIA GTX 1080 over their usual K80 GPU accelerator. It all comes down to single vs double precision and flops/watt computations. Another interesting point is that the team feels that WRF will be their more difficult task, while the other teams believe that Graph500 is the most challenging app.

Youtube Video

Team USTC: The team from the University of Science and Technology of China is a happy team. Their hardware is running smoothly and they're very happy with sponsor Huawei. The team is pushing an eight-node, 320-core cluster with dual NVIDIA K80 GPUs. If you've been following the cluster competitions closely, you'll already know that USTC is a team to watch, having steadily improved in the rankings over the past several years. While they've turned in great performances during parts of the competitions, they have yet to put together a long enough string of application wins to nab them the championship trophy. Maybe this will be their year?

Youtube Video

In our next blog, we'll reveal detailed LINPACK results for the ISC'16 competition and put them into historical context. Stay tuned. ®

Similar topics


Other stories you might like

  • Home-grown Euro chipmaker SiPearl signs deal with HPE, Nvidia
    Claims partnerships will drive development and adoption of exascale computing in Europe

    European microprocessor designer SiPearl revealed deals with Nvidia and HPE today, saying they would up the development of high-performance compute (HPC) and exascale systems on the continent.

    Announced to coincide with the ISC 2022 High Performance conference in Hamburg this week, the agreements see SiPearl working with two big dogs in the HPC market: HPE is the owner of supercomputing pioneer Cray and Nvidia is a leader in GPU acceleration.

    With HPE, SiPearl said it is working to jointly develop a supercomputer platform that combines HPE's technology and SiPearl's upcoming Rhea processor. Rhea is an Arm-based chip with RISC-V controllers, planned to appear in next-generation exascale computers.

    Continue reading
  • AMD nearly doubles Top500 supercomputer hardware share
    Intel loses out as Instinct GPUs power the world’s fastest big-iron system

    Analysis In a sign of how meteoric AMD's resurgence in high performance computing has become, the latest list of the world's 500 fastest publicly known supercomputers shows the chip designer has become a darling among organizations deploying x86-based HPC clusters.

    The most eye-catching bit of AMD news among the supercomputing set is that the announcement of the Frontier supercomputer at the US Department of Energy's Oak Ridge National Laboratory, which displaced Japan's Arm-based Fugaku cluster for the No. 1 spot on the Top500 list of the world's most-powerful publicly known systems.

    Top500 updates its list twice a year and published its most recent update on Monday.

    Continue reading
  • Germany to host Europe's first exascale supercomputer
    Jupiter added to HPC solar system

    Germany will be the host of the first publicly known European exascale supercomputer, along with four other EU sites getting smaller but still powerful systems, the European High Performance Computing Joint Undertaking (EuroHPC JU) announced this week.

    Germany will be the home of Jupiter, the "Joint Undertaking Pioneer for Innovative and Transformative Exascale Research." It should be switched on next year in a specially designed building on the campus of the Forschungszentrum Jülich research centre and operated by the Jülich Supercomputing Centre (JSC), alongside the existing Juwels and Jureca supercomputers.

    The four mid-range systems are: Daedalus, hosted by the National Infrastructures for Research and Technology in Greece; Levente at the Governmental Agency for IT Development in Hungary; Caspir at the National University of Ireland Galway in Ireland; and EHPCPL at the Academic Computer Centre CYFRONET in Poland.

    Continue reading

Biting the hand that feeds IT © 1998–2022