Secrets of the Supercomputers

By Andrew Jones

June 17, 2014

The man pacing the edge of the meeting room was increasingly convinced that at least one, and possibly most, of his colleagues were borderline idiots. He tried again to bring perspective to their latest argument: “So, just to summarize, we’ve spent two hours arguing about only the first page of each bid, have spent less than five minutes passing a few cursory glances through the rest of the hundreds of pages in the oversized bundles the other bidders have dumped on us, and yet we are ruling out this bid because it is only eight pages long in total?”

It had started at noon. Well, no it actually started over a year ago. Typical saga of trying to spend money on anything larger than a pencil in a university environment. But the current meeting, bringing that year to a climax, had only started at noon. At least they had eaten. Although the standard university catering meant not-quite-but-nearly stale “something paste” sandwiches, fruit that had obviously been used as footballs, warm orange juice and half as many cakes as there were attendees.

This was the day that the five members of the CARROT Supercomputer Procurement Project Final Evaluation Team had been both looking forward to and dreading at the same time. The team name changed every time they met because no one knew what the real name was, and the origins of the CARROT acronym had long ago vaporized, but it at least the team title was long enough to sound important.

It was time to open the bids for the supercomputer procurement. The formal request for proposals (RFP) process of the last two months had just closed at noon that day. Parcels containing the bids had arrived by courier or hand delivery all morning. They were now sat in a pile on the table waiting to be opened, carefully read and scored to reveal the winner. Beside the six parcels were a stack of printed copies of the bid evaluation form. Following best practice in supercomputer procurement, this had been written at the same time as the RFP document, based on the desperate belief that comparing the bids against this form would automatically reveal the best bid and ensure a justified solution.

Well, maybe some of them still believed that, but at least two of them had more experience of how procurement evaluations turned out at the nasty end of the process. “Nasty” because the number of unhappy people was sure to be greater than the number of happy people after the end of the meeting.

The team started to tear open the parcels. Each parcel contained five copies of a vendor’s bid, as requested by the procurement process – one for each team member to read. Five bids were vigorous assaults on the rainforests. The vendors had clearly included every bit of information even obliquely relevant to the RFP questions, in the fear that leaving out some obscure detail might cost them the win. One vendor had probably also killed several laser printers in the process, as their bid parcel had needed a goods trolley to get it to the meeting room. The sixth bid was unusual – submitted by a newcomer to the market, it was a mere eight pages, with concise answers for each RFP question.

The team started reading. After a mere handful of minutes, Alfred (director of the hosting supercomputer center) leaned back in his chair and confidently declared that “Red Supercomputers Inc. is obviously the winner”.

What?” exploded the rest of the committee members, as they saw the careful procurement process dissolving into farce before their eyes.

They’ve bid the most cores” Alfred answered. “I’ve read the front page of each bid. Apart from Blue HPC Systems Inc. who didn’t write the number of cores on the first page and so I’ve just ignored them.

Bruce (applications team leader) argued back “but Red Inc. has only bid a quarter of the memory of the others – which explains why they can afford so many cores – but the system will be useless for real applications. Green Supers Inc. have offered a ton of memory though.

Charles (director of the partnering supercomputer center) joined in “Well the winner can’t be Green Supers because they have won the last four procurements I have been involved with and it will start to look like I’m favoring them. Even if it does look like they have bid the best system again.

And then Derek (senior, but role never quite clear) added to the mix with “Yellow HPC Products Inc. have used a silly font – but they have bid nearly as many cores as Red.

Edward (head of the systems group) pushed back from the table, waving the evaluation forms in panic “wait, we haven’t even started the evaluation forms yet.” He was new to this game. Alfred countered this with “we’ll use those forms later, to explain the decision we have made!” The room disintegrated into a cacophony of competing voices. As Edward tried in vain to thrust the evaluation forms under the pens of his colleagues, the path of the meeting had been twisted into a familiar tone that many of them recognized from previous procurements.

As the afternoon rolled on, minute after minute, following each argument circle after circle, the debate centered on Red’s biggest system for the budget vs. Green’s better balanced but smaller system. Bruce folded himself in a depressed heap in the corner, realizing his team would have to sort out the mess of getting users to approve of whatever was deployed and to help make their applications run on it. Edward simply looked on in horror as his (frankly unrealistic) expectations of a carefully planned and rationally conducted evaluation process floated away with the smell of long cold coffee.

Alfred, Charles & Derek soldiered on. Alfred used a calming voice to insert “whatever we buy, however, finely architected for a range of applications, the HPC media, the HPC community and the funding agencies will only look at how many cores or FLOPS we bought. So let’s make sure we get the most – or risk looking like we got a bad deal.

Bruce whimpered from the corner “and what about our users???

Charles said again, as he had already repeated several times that long afternoon, “I really think we should read the rest of Blue’s bid before we make any rash decisions – and consider the evaluation forms we put together.

Alfred shot back: “I’m not reading hundreds of pages of Blue’s proposal to find out what they have actually bid. If they won’t summarize it sharply up front then tough. But I’m not buying from Green either, because I’m not giving all this money to a bid of only eight pages, however good it is!

Derek paused his pacing at the edge of the meeting room and again to bring perspective to their latest argument: “So, just to summarize, we’ve spent two hours arguing about only the first page of each bid, have spent less than five minutes passing a few cursory glances through the rest of the hundreds of pages in the oversized bundles the other bidders have dumped on us, and yet we are ruling out the Green bid because it is only eight pages long in total?”

Derek collapsed back into his chair with a weary battle strained sigh. “I knew we should have got those independent consultants in to help us with this …

These are revelations from inside the strange world of supercomputing centers. Nobody is pretending these are real stories. They couldn’t possibly be. Could they?

NB: No offense to anyone intended. Gentle mocking maybe. Serious lessons definitely.

To Be Continued….

[The serious footnote: buying HPC systems can be a complex task, balancing competing commercial, technical, political and practical drivers.]

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!

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 XL — were added to the benchmark suite as MLPerf continues 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 power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa 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…

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…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen 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…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. 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…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. 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…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

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…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire