RFHPC214: A Look at TOP500 Trends on the Road to Exascale

From left, Henry Newman, Dan Olds, Shahin Khan, and Rich Brueckner are the Radio Free HPC team

In this podcast, the Radio Free HPC team looks at the semi-annual TOP500 BoF presentation by Jack Dongarra.

The TOP500 list of supercomputers serves as a “Who’s Who” in the field of High Performance Computing (HPC). It started as a list of the most powerful supercomputers in the world and has evolved to a major source of information about trends in HPC. The 52nd TOP500 list will be published in November 2018 just in time for SC18. This BoF will present detailed analyses of the TOP500 and discuss the changes in the HPC marketplace during the past years. The BoF is meant as an open forum for discussion and feedback between the TOP500 authors and the user community.

After that, we do our Catch of the Week.

See our complete coverage of SC18

Sign up for our insideHPC Newsletter

Download the MP3 * Subscribe on iTunes * RSS Feed

RFHPC213: Running Down the TOP500 at SC18

In this podcast, the Radio Free HPC team looks back on the highlights of SC18 and the newest TOP500 list of the world’s fastest supercomputers.

Buddy Bland shows off Summit, the world’s fastest supercomputer at ORNL.

The latest TOP500 list of the world’s fastest supercomputers is out, a remarkable ranking that shows five Department of Energy supercomputers in the top 10, with the first two captured by Summit at Oak Ridge and Sierra at Livermore. With the number one and number two systems on the planet, the “Rebel Alliance” vendors of IBM, Mellanox, and NVIDIA stand far and tall above the others.

Summit widened its lead as the number one system, improving its High Performance Linpack (HPL) performance from 122.3 to 143.5 petaflops since its debut on the previous list in June 2018. Sierra also added to its HPL result from six months ago, going from 71.6 to 94.6 petaflops, enough to bump it from the number three position to number two. Both are IBM-built supercomputers, powered by Power9 CPUs and NVIDIA V100 GPUs.

Sierra’s ascendance pushed China’s Sunway TaihuLight supercomputer, installed at the National Supercomputing Center in Wuxi, into third place. Prior to last June, it had held the top position on the TOP500 list for two years with its HPL performance of 93.0 petaflops. TaihuLight was developed by China’s National Research Center of Parallel Computer Engineering & Technology (NRCPC).

In this video from ISC 2018, Yan Fisher from Red Hat and Buddy Bland from ORNL discuss Summit, the world’s fastest supercomputer. Red Hat teamed with IBM, Mellanox, and NVIDIA to provide users with a new level of performance for HPC and AI workloads.

Tianhe-2A (Milky Way-2A), deployed at the National Supercomputer Center in Guangzho, China, is now in the number four position with a Linpack score of 61.4 petaflops. It was upgraded earlier this year by China’s National University of Defense Technology (NUDT), replacing the older Intel Xeon Phi accelerators with the proprietary Matrix-2000 chips.

Top-500, Green-500, IO-500, HPCG, and now CryptoSuper-500 all point to growing versatility of supercomputers,” said Shahin Khan from OrionX. “It’s time to more explicitly recognize that. Counting systems which are capable of doing Linpack but In fact are doing something else continues to be an issue. We need additional info about systems so we can tally them correctly and make this less of a game.”

At number five is Piz Daint, a Cray XC50 system installed at the Swiss National Supercomputing Centre (CSCS) in Lugano, Switzerland. At 21.2 petaflops, it maintains its standing as the most powerful system in Europe. It is powered by a combinations of Intel Xeon processors and NVIDIA Tesla P100 GPUs

Trinity, a Cray XC40 system operated by Los Alamos National Laboratory and Sandia National Laboratories improved its performance to 20.2 petaflops, enough to move it up one position to the number six spot. It uses Intel Xeon Phi processors, the only top ten system to do so.

The AI Bridging Cloud Infrastructure (ABCI) installed in Japan at the National Institute of Advanced Industrial Science and Technology (AIST) is listed at number seven with a Linpack mark of 19.9 petaflops. The Fujitsu-built system is powered by Intel Xeon Gold processors, along with NVIDIA Tesla V100 GPUs.

Germany provided a new top ten entry with SuperMUC-NG, a Lenovo-built supercomputer installed at the Leibniz Supercomputing Centre (Leibniz-Rechenzentrum) in Garching, near Munich. With more than 311,040 Intel Xeon cores and an HPL performance of 19.5 petaflops, it captured the number eight position.

Titan, a Cray XK7 installed at the DOE’s Oak Ridge National Laboratory, and previously the most powerful supercomputer in the US, is now the number nine system. It achieved 17.6 petaflops using NVIDIA K20x GPU accelerators.

Sequoia, an IBM BlueGene/Q supercomputer installed at DOE’s Lawrence Livermore National Laboratory, is the 10th-ranked TOP500 system. It was first delivered in 2011, achieving 17.2 petaflops on HPL.

See our complete coverage of SC18

Sign up for our insideHPC Newsletter

Download the MP3 * Subscribe on iTunes * RSS Feed

RFHPC212: Inside the Spaceborne Supercomputer from HPE

In this podcast, the Radio Free HPC team sits down with Mark Fernandez from HPE to discuss the Spaceborne Supercomputer that it currently orbiting the planet in the International Space Station.

Last week, HPE announced it is opening high-performance computing capabilities to astronauts on the International Space Station (ISS) as part of its continued experiments on the Spaceborne Computer project.

Spaceborne Computer is the first commercial off-the-shelf (COTS) supercomputer that HPE and NASA launched into space for a one-year experiment to test resiliency and performance, achieving one teraFLOP (a trillion floating point operations per second) and successfully operating on the International Space Station (ISS).  After completing its one-year mission proving it can withstand harsh conditions of space – such as zero gravity, unscheduled power outages, and unpredictable levels of radiation – Spaceborne Computer will now, for the first time ever, open its supercomputing capabilities for use aboard the ISS. These “above-the-cloud” services will allow space explorers and experimenters to run analyses directly in space instead of transmitting data to and from Earth for insight.

Our mission is to bring innovative technologies to fuel the next frontier, whether on Earth or in space, and make breakthrough discoveries we have never imagined before,” said Dr. Eng Lim Goh, Chief Technology Officer and Vice President, HPC and AI, HPE. “After gaining significant insights from our first successful experiment with Spaceborne Computer, we are continuing to test its potential by opening up above-the-cloud HPC capabilities to ISS researchers, empowering them to take space exploration to a new level.”

After that, we do our Catch of the Week:

  • Shahin saw some cartoony Beethoven thing that he liked a lot. He also like this video of artists from over 124 countries singing Ghandi’s favorite bhajan, “Vaishnav Jan To Tene Kahiye.”
  • Dan likes the story about the European HPC Handbook. You can’t tell all the players without a program.
    • Dan is starting a 4800 mile road tour on his roundtrip to Dallas for SC18. He’s hitting all the labs and supercomputer centers he can along the way.
  • Rich is proud to announce that his first documentary film has already raised $700 for the Multnomah County Animal Shelter.

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter.

RFHPC211: A Preview of the SC18 Student Cluster Competition

In this podcast, Radio Free HPC Previews the SC18 Student Cluster Competition.

The Student Cluster Competition was developed in 2007 to provide an immersive high performance computing experience to undergraduate and high school students. With sponsorship from hardware and software vendor partners, student teams design and build small clusters, learn designated scientific applications, apply optimization techniques for their chosen architectures, and compete in a non-stop, 48-hour challenge at the SC conference to complete a real-world scientific workload, showing off their HPC knowledge for conference attendees and judges. Teams are composed of six students, at least one advisor, and vendor partners. The advisor provides guidance and recommendations, the vendor provides the resources (hardware and software) and the students provide the skill and enthusiasm. Students work with their advisors to craft a proposal that describes the team, the suggested hardware, and their approach to the competition. The SCC committee reviews each proposal and provides comments for all submissions received before the deadline.”

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter.

RFHPC210: SC18 Preview

In this podcast, the Radio Free HPC team looks ahead to preview SC18 in Dallas. The conference takes place Nov. 11 – 16.

SC18 in Dallas is world’s largest gathering of HPC professionals, and the smart money is on the organizations that leverage the show for their own gatherings, meetups, special booth sessions, and user groups. Here is a roundup of special events not to miss.

Ancillary Events at SC18:

Does your company have special events planned for SC18? Let us know at news @ insidehpc.com and we will add it to this post.

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter.

RFHPC206: A Look at the Frightening NotPetya Cyber Attack

In this podcast, the Radio Free HPC team looks at one of the most massive hacks ever, the NotPetya cyber attack on shipping company Maersk and their partners.

This story, featured in Wired magazine, should send chills down the spines of anyone out there who isn’t religiously updating their machines. In other news, Dan is in Australia for the week at the HPC-AI Advisory Council annual Perth meeting, in his catch of the week, he discusses how one of the companies at the conference has made extensive use of IBM’s Watson and is seeing great benefits. Shahin brings up a new camera with almost unimaginable image specs, while Henry get his two cents in on everything else.”

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter.

RFHPC205: A look at Hot Chips 2018

In this podcast, the Radio Free HPC team looks at the latest developments in processor technology coming out of the recent Hot Chips conference.

Since it started in 1989, HOT CHIPS has been known as one of the semiconductor industry’s leading conferences on high-performance microprocessors and related integrated circuits. The conference is held once a year in August in the center of the world’s capital of electronics activity, Silicon Valley.

The HOT CHIPS conference typically attracts more than 500 attendees from all over the world. It provides an opportunity for chip designers, computer architects, system engineers, press and analysts, as well as attendees from national laboratories and academia to mix, mingle and see presentations on the latest technologies and products. The three days of the conference typically feature two tutorials, two keynotes, a panel discussion and around 25 presentations on a variety of subjects related to microprocessors and integrated circuits conference.

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter.

RFHPC204: A Look at the New Eagle Supercomputer at NREL

In this podcast, the Radio Free HPC team looks at the new Eagle supercomputing under construction at NREL.

The new machine from HPE will run more detailed models that simulate complex processes, systems, and phenomena to advance early research and development on energy technologies across fields including vehicle, wind power, and data sciences.

We are strongly committed to architecting technologies to power the next wave of supercomputing and are creating advanced HPC systems while scaling energy efficiency in data centers, to get us there,” said Bill Mannel, vice president and general manager, HPC and AI Group, HPE. “Through Eagle and our overall ongoing collaboration with the U.S. DOE and NREL, we are advancing research to bolster innovation in energy and sustainability.”

After that, we do our Catch of the Week:

  • Shahin likes the story coming out of DEFCON where 11 year olds were able to hack a local election result web site in just minutes. They had 39 kids from ages 6-17.
  • Dan likes the story about the kid who hacked into Apple and was apprehended with a store of sensitive data. Will they give a job someday?
  • Dan also notes that the new Foreshadow exploit makes your home computer vulnerable to something called SGX. This stuff is never ending, so do we need to rethink how we do computing?
  • Shahin is convinced that Ai is just a subset of HPC.
  • Rich is impressed with NVIDIA’s new Turing GPUs that can do real-time ray tracing. How fast is that? They measure their performance as 10 Gigarays per second.

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter.

RFHPC203: A Look at China’s New Exascale Prototypes

In this podcast, the Radio Free HPC team looks at China’s new ARM-based Exascale supercomputer prototype.

As reported in China Daily, scientists have put an exascale computing prototype into operation that does not run the x86 instruction set. The Sunway exascale computer prototype was developed by the National Research Center of Parallel Computer Engineering and Technology (NRCPC), the team that developed Sunway TaihuLight, crowned the world’s fastest computer two years in a row in 2016 and 2017.

The Sunway exascale computer prototype is very much like a concept car that can run on road,”said Yang Meihong, director of the National Supercomputing Center in Jinan. “We expect to build the exascale computer in the second half of 2020 or the first half of 2021,” said Yang.

Another prototype exascale supercomputer Tianhe-3 passed the acceptance tests on July 22. Its final version is expected to come out in 2020. The two prototypes marked a further step towards China’s successful development of the next-generation supercomputer.

After that, we do our Catch of the Week:

  • Rich points us to the story about Rigetti Computing’s pending 128-qubit quantum computer. The company has already built the 128-qubit processing chip, and is working to put all the pieces together to bring more power to researchers and developers. If successful, it could be the world’s most powerful quantum computer and it could have the chance to outpace traditional supercomputers. Meanwhile, you can already access IBMQ, D-Wave, and Chinese quantum machines in the cloud today.
  • Henry notes that one of the scary things to come out of the Black Hat conference is a new kind of microwave weapon for cooking your enemy from communication satellites.
  • Dan is wardriving in North Carolina in the town the Internet forgot.

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter.

RFHPC202: A look at the IO500 Benchmark Suite with John Bent

In this podcast, the Radio Free HPC team talks to John Bent from the IO500 committee about why he and a team of I/O professionals created the IO500 benchmark suite. The second IO500 list was revealed at ISC 2018 in Frankfurt, Germany.

Following the success of the Top500 in collecting and analyzing historical trends in supercomputer technology and evolution, the IO500 was created in 2017 and published its first list at SC17. The need for such an initiative has long been known within High Performance Computing; however, defining appropriate benchmarks had long been challenging. Despite this challenge, the community, after long and spirited discussion, finally reached consensus on a suite of benchmarks and a metric for resolving the scores into a single ranking.

The multi-fold goals of the benchmark suite are as follows:

  1. Maximizing simplicity in running the benchmark suite
  2. Encouraging complexity in tuning for performance
  3. Allowing submitters to highlight their “hero run” performance numbers
  4. Forcing submitters to simultaneously report performance for challenging IO patterns.

Specifically, the benchmark suite includes a hero-run of both IOR and mdtest configured however possible to maximize performance and establish an upper-bound for performance. It also includes an IOR and mdtest run with highly prescribed parameters in an attempt to determine a lower-bound. Finally, it includes a namespace search as this has been determined to be a highly sought-after feature in HPC storage systems that has historically not been well-measured. Submitters are encouraged to share their tuning insights for publication.

Once again, we encourage you to submit (see http://io500.org/submission), to join our community, and to attend our BoF “The IO-500 and the Virtual Institute of I/O” at ISC 2018 where we will announce the second ever IO500 list. The current list includes results from BeeGFS, DataWarp, IME, Lustre, and Spectrum Scale. We hope that the next list has even more.

In this video from ISC 2018, John Bent and Jay Lofstead describe how the IO500 benchmark measures storage performance in HPC environments.

After that, we do our Catch of the Week:

  • Shahin notes that Apple has become the first Trillion-dollar company. He was also on a AI panel where they talked about how the HR department may soon become the Human and Machine Resources Department.
  • Henry points us to the story that DEFCON conference is hosting a contest for 8 year olds to hack government election sites.
  • Rich is taunting Dan with the impressive specifications on his new I9-powered MacBook Pro. There will be a render-off.
  • John Bent went to Google IO and says they were very excited about their new Gmail feature that will auto-compose emails for you. Rich wants to use it on illiterate PR people who think they can call him at 4 am.
  • Dan likes the story about the man who ran out of data on his phone and woke a sleeping couple in Palo Alto to demand they give him their WiFi password. The incident started a crime spree, as the man stole their kid’s bicycle to flee the scene.

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter.