Exascale: Power Is Not the Problem!

By Andrew Jones

August 29, 2011

To build exascale systems, power is probably the biggest technical hurdle on the hardware side. In terms of getting to exascale computing, demonstrating the value of supercomputing to funders and the public is a more urgent challenge. But the top roadblock for realizing the potential benefits from exascale is software.

That title is probably controversial to most readers. It is likely that if you asked members of the supercomputing community what is the single biggest challenge for exascale computing, the most common answer would be “power.” It is widely reported, widely talked about, and in many places, generally accepted that finding a few orders of magnitude improvement in power consumption is the biggest roadblock on the way to viable exascale computing. Otherwise, the first exascale computers will require 60MW, 120MW or 200MW — pick your favorite horror figure. I’m not so convinced.

I’m not saying the power estimates for exascale computing are not a problem — they are — but they are not the problem. Because, in the end, it is just a money problem. For most in the community, the objection is not so much to the fact of 60-plus MW supercomputers. Instead, the objection is the resulting operating costs of 60-plus MW supercomputers. We simply don’t want to pay $60 million each year for electricity (or more precisely we don’t want to have to justify to someone else — e.g., funding agencies — that we need to pay that much). But why are we so concerned about large power costs?

Are we really saying, with our concerns over power, that we simply don’t have a good enough case for supercomputing — the science case, business case, track record of innovation delivery, and so on? Surely if supercomputing is that essential, as we keep arguing, then the cost of the power is worth it.

There are several large scientific facilities that have comparable power requirements, often with much narrower missions — remember that supercomputing can advance almost all scientific disciplines — for example, LHC, ITER, NIF, and SNS. And indeed, most of the science communities behind those facilities are also large users of supercomputing.

I occasionally say, glibly and deliberately provocatively, if the scientific community can justify billions of dollars, 100MW of power, and thousands of staff in order to fire tiny particles that most people have never heard of around a big ring of magnets for a fairly narrow science purpose that most people will never understand, then how come we can’t make a case for a facility needing only half of those resources that can do wonders for a whole range of science problems and industrial applications?

[There is a partial answer to that, which I have addressed on my HPC Notes blog to avoid distraction here.]

But secondly, and more importantly, the power problem can be solved with enough money if we can make the case. Accepting huge increases in budgets would also go a long way toward solving several of the other challenges of exascale computing. For example, resiliency could be substantially helped if we could afford comprehensive redundancy and other advanced RAS features; data movement challenges could be helped if we could afford huge increases in memory bandwidth at all levels of the system; and so on.

Those technical challenges would not be totally solved but they would be substantially reduced by money. I don’t mean to trivialize those technical challenges, but certainly they could be made much less scary if we weren’t worried about the cost of solutions.

So, the biggest challenge for exascale computing might not be power (or your other favorite architectural roadblock) but rather our ability to justify enough budget to pay for the power, or more expensive hardware, etc. However, beyond even that, there is a class of challenges for which money alone is not enough.

Assume a huge budget meant an exascale computer with good enough resiliency, plenty of memory bandwidth and every other needed architectural attribute was delivered tomorrow, and never mind the power bills. Could we use it? No. Because of a series of challenges that need not only money, but also lots of time to solve, and in most cases need research because we just don’t know the solutions.

I am thinking of the software related challenges.

Even if we have highly favorable architectures (expensive systems with lots of bandwidth, good resiliency, etc.) I think the community and most, if not all, of the applications are still years away from having algorithms and software implementations that can exploit that scale of computing efficiently.

There is a reasonable effort underway to identify the software problems that we might face in using exascale computing (e.g., IESP and EESI). However, in most cases we can only identify the problems; we still don’t have much idea about the solutions. Even where we have a good idea of the way forward, sensible estimates of the effort required to implement software capable of using exascale computing — OS, tools, applications, post-processing, etc. — is measured in years with large teams.

It certainly requires money, but it needs other scarce resources too, specifically time and skills. That involves a large pool of skilled parallel software engineers, scientists with computational expertise, numerical algorithms research and so on. Scarce resources like these are possibly even harder to create than money!

Power is a problem for exascale computing, and with current budget expectations is probably the biggest technical challenge for the hardware. In terms of getting to exascale computing, demonstrating the value of increased investment in supercomputing to funders and the public/media is probably a more urgent challenge. But the top roadblock for achieving the hugely beneficial potential output from exascale computing is software. There are many challenges to do with the software ecosystem that will take years, lots of skilled workers, and sustained/predictable investment to solve.

That “sustained/predictable” is important. Ad-hoc research grants are not an efficient way to plan and conduct a many-year, many-person, community-wide software research and development agenda. Remember that agenda will consume a non-trivial portion of the careers of many of the individuals involved. And when the researchers start out on this necessary software journey, they need confidence that funding will be there all the way to production deployment and ongoing maintenance many years into the future.

About the Author

Andrew is Vice-President of HPC Services and Consulting at the Numerical Algorithms Group (NAG). He was originally a researcher using HPC and developing related software, later becoming involved in leadership of HPC services. He is also interested in exascale, manycore, skills development, broadening usage, and other future concerns of the HPC community.

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