A Beginner’s Guide to the ASC19 Finals

By Alex Woodie

April 22, 2019

Three thousand watts. That’s how much power the competitors in the 2019 ASC Student Supercomputer Challenge here in Dalian, China, have to work with. Everybody would like more juice to run compute-intensive HPC simulations and AI models, which means that whoever can get the most work done within the power constraints will take home the crown.

Twenty student teams from around the world have gathered here at the Dalian University of Technology to pit their computing abilities against one another in a no-holds-barred computational extravaganza. The event, which is hosted by the Asia Supercomputer Community (ASC) and sponsored by Inspur and Intel, started on Sunday and runs through Thursday, when the winners will be announced.

Just getting to this point is a victory of sorts for the undergraduates, but everybody would like to take home an award just the same. The ASC19 competition opened for registration in November with 300 student teams from 200 universities. The preliminary round kicked off in January and continued until early March, when the teams submitted their final proposals. A jury of HPC and AI experts with the ASC organization judged the proposals, and selected the final teams that are here this week.

The first two days of ASC (Sunday and Monday) are all about building the cluster, completing test runs, and dialing in the cluster for the actual competition, which runs mainly Tuesday and Wednesday. The configuration phase is a critical period for the teams, as they must determine which configuration will deliver the best performance against the various applications.

Whirl-Win

The sound of whirring fans fills the spacious indoor stadium here whenever one of the teams starts a test and the clusters ramp up. The teams check the results to see how the cluster ran, make some hardware or software tweaks, and repeat as necessary until they’re happy (or at least somewhat satisfied) with the setup.

Twenty student teams from around the world compete in ASC19 this week at Dalian University of Technology’s indoor stadium

The big challenge is to pick a cluster configuration that gives good results across all of the applications, or at least a cross-section of them. Dialing in that configuration involves balancing numerous variables, some of which are known and some of which are not. Once competition begins, the teams are not allowed to re-boot their clusters, although they can take action such as idling spare CPU or GPU nodes, which gives them some flexibility to adapt to different workloads while staying under 3,000 watts.

Each team is allowed to use as many servers as it wants. Boxes of Inspur servers are lined up against the bleacher seats, waiting to be racked and put into action. A maximum of 384GB of RAM is allowed, which most teams will probably use, and teams must supply their own hard drives (SSDs are a must).

The teams are allowed to use any internal networks they like. Most have chosen Mellanox interconnects, but one team has selected Intel OmniPath. Will it provide an advantage? Only time will tell.

ASC19 teams are also allowed to bring as many accelerators as they want to the competition. One team is rumored to have a large number of Nvidia V100 GPUs at the ready — effectively “loaded for bear.” But because of that pesky 3,000 watt limit, it’s unclear if the ASC19 winner will able to bring all that firepower to bear.

Strategic Configs

Some of the applications will clearly benefit from GPU computational power, while others will favor a lighter-weight approach. It all comes down to finding a strategic, or balanced, approach.

The power board shows how close the 20 teams competing in ASC19 are coming to the 3,000-watt max

For example, one of the ASC19 applications is the single-image super-resolution (SISR) competition, which requires the teams to use the PyTorch deep learning framework to basically “up-convert” a series of blurry images into images with greater detail. Having a lot of GPUs here could be a boon.

Another prominent application that’s sure to stress the ASC19 teams is the Community Earth System Model (CESM), which is one of the main models used by the United Nations Intergovernmental Panel on Climate Change (IPCC). CESM is more of a “classic” HPC application, which may or may not benefit from GPUs.

The teams are eagerly awaiting the release of the actual data sets on Tuesday, which will go far in determining whether they made the right decisions in configuring their setups. Of course, by then it will be too late to make any major changes to their configuration.

Besides the CESM and the SISR tests, the teams must also prepare for other applications, including a WTDBG sequence assembly application; the HPL & HPCG high performance benchmarks; and the old standby benchmark, LINPACK. There is also one mystery application that will test the teams’ ability to adapt to the unexpected.

Except for the mystery application, the ASC19 competitors have known what applications, algorithms, frameworks, and codes they’re going to be working with here. That allows them to prepare, including by writing CUDA programs for AI and HPC applications. That’s why the release of the datasets early Tuesday morning is such a critical event. The data sets could be significantly bigger than what they’re used to dealing with, which could throw an interesting wildcard into the mix.

What’s At Stake

The computational tests take place Tuesday and Wednesday, and on Thursday morning the teams will make their presentation and mount a verbal defense against the jury. Later that day, the winners will be announced and a banquet will be held in celebration.

The prizes at play include the title of Champion, the Silver Prize, the Highest Linpack, the e-Prize, and the Group Competition, among other awards. There’s a total bonus of over $36,000 at stake.

As the students battle it out on the stadium floor for AI and HPC supremacy, there will be other events taking place, including the 21st HPC Connection Workshop and Artificial Intelligence & Supercomputing Innovation Symposium.

Speaking at the workshop will be Jack Dongarra, the chairman of the ASC Advisory Committee and a distinguished professor at the University of Tennessee, as well as University of California Berkeley expert Leon Chua.

The symposium, meanwhile, will feature Huchuan Lu, a professor at Dalian University of Technology, and Guohui Li, a researcher at the Dalian Institute of Chemical Physics in the Chinese Academy of Sciences.

The main event, though, involves the competition among the 300-plus students from around the world. The students hail from 15 universities in China and five from universities in other countries, including Germany (two), Poland, South Korea, and Columbia. Universities represented include:

  • Beihang University, China

    Boxes of Inspur servers line the wall of the ASC19 Finals this week in Dalian, China
  • Southern University of Science and Technology, China
  • Fuzhou University, China
  • Jinan University, China
  • Shanghai Jiao Tong University, China
  • Sun Yat-Sen University, China
  • Taiwan Tsing Hua University, China
  • Peking University, China
  • Huazhong University of Science and Technology, China
  • Shanxi University, China
  • Taiyuan University of Technology, China
  • University of Electronic Science and Technology of China, China
  • The Chinese University of Hong Kong, China
  • University of Warsaw / Warsaw University of Technology, Poland
  • Tsinghua University, China
  • Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany
  • University of Tartu, Germany
  • Universidad EAFIT, Columbia
  • Sungkyunkwan University, South Korea
  • Dalian University of Technology, China

Stay tuned to HPCwire this week for more news from Dalian as ASC19 unfolds.

 

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

Quantum Software Specialist Q-CTRL Inks Deals with IBM, Rigetti, Oxford, and Diraq

September 10, 2024

Q-CTRL, the Australia-based start-up focusing on quantum infrastructure software, today announced that its performance-management software, Fire Opal, will be natively integrated into four of the world's most advanced qu Read more…

Computing-Driven Medicine: Sleeping Better with HPC

September 10, 2024

As a senior undergraduate student at Fisk University in Nashville, Tenn., Ifrah Khurram's calculus professor, Dr. Sanjukta Hota, encouraged her to apply for the Sustainable Research Pathways Program (SRP). SRP was create Read more…

LLNL Engineers Harness Machine Learning to Unlock New Possibilities in Lattice Structures

September 9, 2024

Lattice structures, characterized by their complex patterns and hierarchical designs, offer immense potential across various industries, including automotive, aerospace, and biomedical engineering. With their outstand Read more…

NSF-Funded Data Fabric Takes Flight

September 5, 2024

The data fabric has emerged as an enterprise data management pattern for companies that struggle to provide large teams of users with access to well-managed, integrated, and secured data. Now scientists working at univer Read more…

xAI Colossus: The Elon Project

September 5, 2024

Elon Musk's xAI cluster, named Colossus (possibly after the 1970 movie about a massive computer that does not end well), has been brought online. Musk recently posted the following on X/Twitter: "This weekend, the @xA Read more…

Researchers Benchmark Nvidia’s GH200 Supercomputing Chips

September 4, 2024

Nvidia is putting its GH200 chips in European supercomputers, and researchers are getting their hands on those systems and releasing research papers with performance benchmarks. In the first paper, Understanding Data Mov Read more…

Quantum Software Specialist Q-CTRL Inks Deals with IBM, Rigetti, Oxford, and Diraq

September 10, 2024

Q-CTRL, the Australia-based start-up focusing on quantum infrastructure software, today announced that its performance-management software, Fire Opal, will be n Read more…

NSF-Funded Data Fabric Takes Flight

September 5, 2024

The data fabric has emerged as an enterprise data management pattern for companies that struggle to provide large teams of users with access to well-managed, in Read more…

Shutterstock 1024337068

Researchers Benchmark Nvidia’s GH200 Supercomputing Chips

September 4, 2024

Nvidia is putting its GH200 chips in European supercomputers, and researchers are getting their hands on those systems and releasing research papers with perfor Read more…

Shutterstock 1897494979

What’s New with Chapel? Nine Questions for the Development Team

September 4, 2024

HPC news headlines often highlight the latest hardware speeds and feeds. While advances on the hardware front are important, improving the ability to write soft Read more…

Critics Slam Government on Compute Speeds in Regulations

September 3, 2024

Critics are accusing the U.S. and state governments of overreaching by including limits on compute speeds in regulations and laws, which they claim will limit i Read more…

Shutterstock 1622080153

AWS Perfects Cloud Service for Supercomputing Customers

August 29, 2024

Amazon's AWS believes it has finally created a cloud service that will break through with HPC and supercomputing customers. The cloud provider a Read more…

HPC Debrief: James Walker CEO of NANO Nuclear Energy on Powering Datacenters

August 27, 2024

Welcome to The HPC Debrief where we interview industry leaders that are shaping the future of HPC. As the growth of AI continues, finding power for data centers Read more…

CEO Q&A: Acceleration is Quantinuum’s New Mantra for Success

August 27, 2024

At the Quantum World Congress (QWC) in mid-September, trapped ion quantum computing pioneer Quantinuum will unveil more about its expanding roadmap. Its current Read more…

Everyone Except Nvidia Forms Ultra Accelerator Link (UALink) Consortium

May 30, 2024

Consider the GPU. An island of SIMD greatness that makes light work of matrix math. Originally designed to rapidly paint dots on a computer monitor, it was then Read more…

Atos Outlines Plans to Get Acquired, and a Path Forward

May 21, 2024

Atos – via its subsidiary Eviden – is the second major supercomputer maker outside of HPE, while others have largely dropped out. The lack of integrators and Atos' financial turmoil have the HPC market worried. If Atos goes under, HPE will be the only major option for building large-scale systems. Read more…

AMD Clears Up Messy GPU Roadmap, Upgrades Chips Annually

June 3, 2024

In the world of AI, there's a desperate search for an alternative to Nvidia's GPUs, and AMD is stepping up to the plate. AMD detailed its updated GPU roadmap, w Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst fir Read more…

Shutterstock_1687123447

Nvidia Economics: Make $5-$7 for Every $1 Spent on GPUs

June 30, 2024

Nvidia is saying that companies could make $5 to $7 for every $1 invested in GPUs over a four-year period. Customers are investing billions in new Nvidia hardwa Read more…

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l Read more…

Shutterstock 1024337068

Researchers Benchmark Nvidia’s GH200 Supercomputing Chips

September 4, 2024

Nvidia is putting its GH200 chips in European supercomputers, and researchers are getting their hands on those systems and releasing research papers with perfor Read more…

Leading Solution Providers

Contributors

IonQ Plots Path to Commercial (Quantum) Advantage

July 2, 2024

IonQ, the trapped ion quantum computing specialist, delivered a progress report last week firming up 2024/25 product goals and reviewing its technology roadmap. Read more…

Intel’s Next-gen Falcon Shores Coming Out in Late 2025 

April 30, 2024

It's a long wait for customers hanging on for Intel's next-generation GPU, Falcon Shores, which will be released in late 2025.  "Then we have a rich, a very Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Department of Justice Begins Antitrust Probe into Nvidia

August 9, 2024

After months of skyrocketing stock prices and unhinged optimism, Nvidia has run into a few snags – a  design flaw in one of its new chips and an antitrust pr Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

MLPerf Training 4.0 – Nvidia Still King; Power and LLM Fine Tuning Added

June 12, 2024

There are really two stories packaged in the most recent MLPerf  Training 4.0 results, released today. The first, of course, is the results. Nvidia (currently Read more…

Spelunking the HPC and AI GPU Software Stacks

June 21, 2024

As AI continues to reach into every domain of life, the question remains as to what kind of software these tools will run on. The choice in software stacks – Read more…

Quantum Watchers – Terrific Interview with Caltech’s John Preskill by CERN

July 17, 2024

In case you missed it, there's a fascinating interview with John Preskill, the prominent Caltech physicist and pioneering quantum computing researcher that was Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire