ASC24 Student Cluster Competition: Who Won and Why?

By Dan Olds

June 18, 2024

As is our tradition, we’re going to take a detailed look back at the recently concluded the ASC24 Student Cluster Competition (Asia Supercomputer Community) to see not only who won the various awards, but to figure out how they did it.

This also gives us a chance to see how the various teams did on specific applications and highlight the teams that did well, but maybe didn’t take home a major trophy.

Let’s start at the beginning with the benchmarks:

With an average score of 6.83 points and a high score of 9, HPL wasn’t much of a challenge. Team Zhejiang notched an early win, and why not? They’ve won more ASC Highest LINPACK awards than any other ASC team with this, their third trophy.

Zhejiang won comfortably vs. second place Sun-Yat Sen, but third place USTC missed out on second by only 0.03 of a point, ouch! Shanxi almost made the top three, lagging USTC by a mere 0.07 of a point. Ouch again. Peking secured the fifth-place position just to stay in the conversation.

This is a little unusual. When a team wins LINPACK, the embarrassingly parallel benchmark that shows the very best performance you will see from your cluster, they usually don’t also win HPCG, which will typically show the worst side of your system. Well, Team Zhejiang, broke that mold by taking first in both benchmarks and it wasn’t all that close. Sun-Yat Sen tried to make a race of it but finished nearly .75 points behind.

USTC grabbed third, holding off Team Peking by half a point. Beihang nearly took fourth, missing out by the narrowest margin possible – 0.01 of a point. The average/median HPCG score 5.02/6.35, and the high score was 9, so it was another lay-up for most teams.

Now it gets serious. There was a lot of separation on the LLM Inference challenge. The high score was 14.23 and the average/median came in at 3.13/ and, wait for it, 0.00. This means a lot of teams couldn’t turn in a valid LLM score, so this was a tougher application than most observers thought it would be.

Team Zhejiang is putting together a real run with their third consecutive victory – and by a wide margin. If they can keep this up, they’re a shoe in for the coveted ASC24 championship trophy. Shanghai Jiao Tong makes their first appearance in the leaderboard, but they were a little more than two points behind Zhejiang.

We see Shanxi made a move into third, but came up two points short of overtaking second place. Team Peking is still hanging around in fourth and Sun-Yat Sen in fifth, neither within shouting distance of the top three finishers.

End of Day 1, Now for Day 2

After what was probably a restless night of  tossing and turning, the 25 teams return to take on another long day of running apps and fixing problems. This is going to be the hardest day of the competition. There are no more easy workloads, they’re all hard from here on in.

GoMars was, judging by the results, the easiest workload of the day. But this doesn’t mean it was easy. The average/median score was only 5.31/2.81 points – that’s quite a negative skew. Shanghai Jiao Tong sets the bar with the high score at 16.92.

Team Sun-Yat Sen tried to reel them in but missed the mark by about .75 of a point. Peking moves up into third, finishing about 1.5 points behind Sun Yat-Sen. Harbin can’t get there, trailing the top three by a couple of points. In fifth, we see a new team, SW Petroleum University, posting a score that was less than a full point below Harbin – nice work Team Oil & Gas!!

Now for what everyone thought was going to be the most difficult application of all:  OpenCAEPoro. The average score was 4.12 and the median came in at 3.0, a positive skew, meaning there were some outliers on the high side.

Taiwan’s National Tsing Hua University made a statement with their top score of 15.75 points. Sun-Yat Sen isn’t too far behind, less than a point, Team Lanzhou hits the board for the first time in third, but is two points and change behind the leaders. USTC and new to the board Hong Kong Polytechnic lagged in fourth and fifth place.

The Mystery Application was more than a mystery – it was a horror show. The average score was only 2.98 and the median was just over one point. Why so low? Two major reasons come to mind. First, the students only learn about this application on Day 2, giving them no way to prepare for it. Second? Who the hell has ever heard of WannierTools? I spent time researching it and there isn’t much of anything on the web. Unless you have some serious solid-state physics chops, you’ve probably never heard of WannierTools, Wannier Functions, or even Wannier himself.

But…the nature of this competition is to seriously challenge the student competitors. Selecting WanierTools was a brilliant way to do this, it really threw a curve at them. In most cluster competitions, the mystery application is something a bit more well known and a bit better documented. By picking WannierTools, the organizers posed a hugely difficult task for the students.

Team Peking was up to the task, however, and they hit WannierTools out of the park with the top score of 16 points. Beihang was a distant second by nearly five full points, Shanghai Jiao Tong and Team Jinan were three and four points behind Beihang respectively. But a bright spot is Quinghai University, who pulled into fifth place, and onto our leaderboard, finishing ahead of 20 other teams. Nice job, Team Qinghai!

That was a very tough second day but students have to put it behind them and prepare for their judges interview. I’m betting there were a lot of lights and laptops on well into the evening as students either started, or hopefully, put the finishing touches on their presentations.

The teams did a good job on their presentations, phew. The average score was 7.4 and the median was 7.38, so most of them did well.

Team Peking took home top presentation honors with their score of 9.63 in front of a tough panel of judges. Sun Yat-Sen grabbed second place but was off the pace by almost a full point. Jinan was just fractions of a point behind second place with Zhejiang another half point behind in fourth. Beihang finishes out the top five, a mere 0.04 points behind Zhejiang.

Final Finish!

This was a brutally difficult competition. To me, it was the toughest yet. But, wow, the students came a long, long way in terms of learning about HPC/AI and rising to the occasion while under a lot of pressure. This experience is going to pay dividends to them for the rest of their lives.

Team Peking takes home the ASC24 Championship with an overall score of 69.77 points. Sun Yat-Sen was only six points behind in second. Shanghai Jiao Tong was less than five points behind Sun Yat-Sen in third place. Zhejiang, a team that started very strong, landed in fourth place with Shanxi taking fifth – their highest place ever.

So What Have We Learned?

Looking back on ASC24, a few things come to mind. First, the ASC organizers are, well, really organized. There weren’t any problems that I could see and the students were very well cared for and all enjoyed the experience.

Second, the ASC competitions are tough.

Third, who the hell has ever heard of WannierTools?

It was a great competition and has me looking forward to ASC25. Also, let’s get some more western teams into this event. Students will get a whole new perspective on HPC and AI, coming back with solid real-world skills that will make them stand out from the crowd. But, more importantly, it’s an unforgettable experience in all the best ways.

I’m already looking forward to ASC25…..


Here’s something I found at a student grocery store at Shanhai University. It was sitting quietly among the array of cookies and packaged pastries on a shelf. Its official no nonsense look caught my attention and it’s sheer density when I picked it up intrigued me. The name got me too:  Energy Accumulation Compressed Biscuit. While descriptive, it tells me nothing about what I’m buying into here. This only heightened my interest.

I instantly dropped a package into my basket of hotel rations, definitely feeling the added weight. It sat on the dresser in my room for the next six nights and then was packed into my luggage for the long trip home. It was so dense that I worried that it might damage other things in my suitcase or even cause me to go over my allotted weight allowance with the airline.

At home, it sat near my suitcase unnoticed for about a week, buried in camera equipment. I finally uncovered it and started my investigation.

It weighs in at just over a pound, which is more what you’d expect from something that size if it were made of metal or stone, not compressed biscuit material. The back panel gives a hint of the potential inside the package with numbers indicating it yields solid percentages of “NRV”, which I assume is some daily requirement of something.

On opening it up, I placed it on the scale and confirmed the weight – a pound plus.

Inside the box, I found eight highly dense obelisks of highly compressed energy. They smelled a bit like graham crackers, which was reassuringly familiar as my next task was the taste test.

It was easy to bite off a chunk. I started with one third of a bar and gingerly started chewing. The first thing that flashed into my mind was “not bad, decent taste” closely followed by “I didn’t know anything could be this dry.”

After swallowing it, aided by a full glass of water, I could feel something like the equivalent of half a sandwich in my stomach. It was a good feeling to have consumed that much with only a single bite of one of the bars – no sandwich preparation, no implements to wash, no plates, no residue. Plus, a flush of energy for the rest of my day.

That sample led to more tastings and I’m now a big fan of the Energy Accumulation Compresses Biscuit. I did some more investigation and found that the “Apocalypse Equipped” blog gave it high marks. That’s good enough for me!

So on my next trip to China, I’m not only getting more of these, I’m going to make sure I can get them the way they were meant to be enjoyed – served up in a study metal box!

Just one more thing to look forward to at ASC25, right? The joys of travel!

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!

ARM, Fujitsu Targeting Open-source Software for Power Efficiency in 2-nm Chip

July 19, 2024

Fujitsu and ARM are relying on open-source software to bring power efficiency to an air-cooled supercomputing chip that will ship in 2027. Monaka chip, which will be made using the 2-nanometer process, is based on the Read more…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI projects.). Other tools have joined the CUDA castle siege. AMD 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 recently posted by CERN’s department of experimental physi Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to communities across the globe. As climate change is warming ocea Read more…

Researchers Say Memory Bandwidth and NVLink Speeds in Hopper Not So Simple

July 15, 2024

Researchers measured the real-world bandwidth of Nvidia's Grace Hopper superchip, with the chip-to-chip interconnect results falling well short of theoretical claims. A paper published on July 10 by researchers in the U. Read more…

Belt-Tightening in Store for Most Federal FY25 Science Budets

July 15, 2024

If it’s summer, it’s federal budgeting time, not to mention an election year as well. There’s an excellent summary of the curent state of FY25 efforts reported in AIP’s policy FYI: Science Policy News. Belt-tight Read more…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI pro Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to com Read more…

Shutterstock 1886124835

Researchers Say Memory Bandwidth and NVLink Speeds in Hopper Not So Simple

July 15, 2024

Researchers measured the real-world bandwidth of Nvidia's Grace Hopper superchip, with the chip-to-chip interconnect results falling well short of theoretical c Read more…

Shutterstock 2203611339

NSF Issues Next Solicitation and More Detail on National Quantum Virtual Laboratory

July 10, 2024

After percolating for roughly a year, NSF has issued the next solicitation for the National Quantum Virtual Lab program — this one focused on design and imple Read more…

NCSA’s SEAS Team Keeps APACE of AlphaFold2

July 9, 2024

High-performance computing (HPC) can often be challenging for researchers to use because it requires expertise in working with large datasets, scaling the softw Read more…

Anders Jensen on Europe’s Plan for AI-optimized Supercomputers, Welcoming the UK, and More

July 8, 2024

The recent ISC24 conference in Hamburg showcased LUMI and other leadership-class supercomputers co-funded by the EuroHPC Joint Undertaking (JU), including three Read more…

Generative AI to Account for 1.5% of World’s Power Consumption by 2029

July 8, 2024

Generative AI will take on a larger chunk of the world's power consumption to keep up with the hefty hardware requirements to run applications. "AI chips repres Read more…

US Senators Propose $32 Billion in Annual AI Spending, but Critics Remain Unconvinced

July 5, 2024

Senate leader, Chuck Schumer, and three colleagues want the US government to spend at least $32 billion annually by 2026 for non-defense related AI systems.  T 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…

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…

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…


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…

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…

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…

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…

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…

Leading Solution Providers


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…

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…

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…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

  • arrow
  • Click Here for More Headlines
  • arrow