HPC and Science & Technology Policy

By Gary M. Johnson

November 26, 2013

Policies, like sausages, cease to inspire respect in proportion as we know how they are made. This is a loose paraphrase of a common analogy. Is it apt when applied to science and technology policy? In particular, how – and where – is HPC policy made? Is it currently being made at all?

Given the number of HPC issues currently on the table and in competition for scarce federal resources, perhaps we should set aside any lack of respect for the process and get engaged in the sausage making. We need clear federal policy guidance to allocate resources and mobilize the technology agencies to advance HPC and its applications. Some suggestions for HPC community action to improve this situation are included.

Moments in S&T Policy

To put the current S&T policy situation in context and to help frame issues, consider this (very) short list of notable S&T policy moments. Note that they are all linked to global conflicts.

National Advisory Committee for Aeronautics (NACA) – 3 March 1915

NACA, the progenitor of NASA, was created during World War I as a reaction to the advances being made in Europe in the new field of aeronautics. It was feared that the US would lose its early lead in aeronautics and that this would have not only military but economic consequences. Consequently, the US created a federally funded organization whose core mission embraced cooperation with industry. This event, unique in our history, is reported to have “slipped through” the legislative process “almost unnoticed as a rider attached to the Naval Appropriation Bill, on 3 March 1915”.  As they say, the rest is history.  NACA/NASA, working in cooperation with the US aeronautics industry, yielded fantastic results.

Science: The Endless Frontier – 25 July 1945

World War II saw the next surge in science and technology. Adding to the laboratories operated by NACA, many further elements of our current federal R&D infrastructure were put in place. At the end of the war, Vannevar Bush, who headed the Office of Scientific Research and Development (OSRD) – which coordinated almost all wartime military R&D, including initiation and early administration of the Manhattan Project – produced the seminal S&T policy document entitled Science: The Endless Frontier. It contained a very interesting “program for action”.  Here are a few excerpts (emphasis ours):

The Government should accept new responsibilities for promoting the flow of new scientific knowledge and the development of scientific talent in our youth. These responsibilities are the proper concern of the Government, for they vitally affect our health, our jobs, and our national security.  …

The effective discharge of these new responsibilities will require the full attention of some over-all agency devoted to that purpose.  

On the wisdom with which we bring science to bear in the war against disease, in the creation of new industries, and in the strengthening of our Armed Forces depends in large measure our future as a nation.

Bush, who was in effect the country’s first presidential science advisor, didn’t get everything he asked for.   No “over-all agency” was created, but the case he made did lead to the creation of the National Science Foundation and Science: The Endless Frontier has constituted the S&T policy basis for all subsequent developments.

In the wrangling that followed Bush’s recommendations, we got a civilian agency focused on supporting university-based research and, to a lesser extent, education.  The military proceeded independently to enlarge its research complex.  Any expansion to other agencies of the NACA-style collaboration with civilian industry fell through the cracks.

Sputnik – 4 October 1957

Our final S&T policy moment is drawn from the Cold War period.  When the Soviet Union launched the first artificial Earth satellite, Sputnik, the event sent shock waves through the US public.  Government actions, spurred on by public reaction to the Sputnik crisis, led to:

So, thanks to the Cold War and Sputnik, we got: a big boost to S&T related education; an expansion of military R&D; and the extension of NACA-style collaboration with industry into the new NASA and for peaceful uses of space.

Beyond NASA, civilian technology agencies, in large measure, still lacked any core mission elements authorizing – and thus protecting – direct collaborations with industry.  This ends our story and is basically the end of significant S&T policy developments in the US.

Today, we have a national S&T policy that hasn’t changed much since the early days of the Cold War.  Perhaps its biggest shortcoming is that, beyond NASA, it doesn’t really empower strong collaborations between federal S&T agencies and American industries.  Inside the Capital Beltway, mission trumps all other considerations.  If S&T agencies engage in activities without direct and obvious linkages to their core missions, they may endanger their budget appropriations.

No agency has the mission to advance HPC.  Several agencies engage in HPC R&D to advance other mission elements, but no agency has advancing HPC itself (like NASA advances aeronautics and space technologies) as a core mission element.

S&T Policy Players

S&T policy is a game that anyone can play – and many do.  Thus, there is no shortage of literature on this topic and there are lots of players, both inside and outside the federal government.  So that your eyes don’t glaze over, we’ll limit the discussion here to mentioning just a few of the main federal players.  We’ve included a short appendix at the end of this article containing some additional information about the players.  Please consult it to make this section less cryptic.  For an overview of S&T policymaking, consult this Primer produced by the Congressional Research Service.

Executive Branch

In the Executive Branch, the main policy players are the Office of Science and Technology Policy (OSTP) and the Office of Management and Budget (OMB).  These offices are part of the Executive Office of the President (EOP).  The OSTP and OMB directors jointly issue an annual memo to federal agencies on science and technology priorities.  The memo on Science and Technology Priorities for the FY 2015 Budget is an interesting read for exascale computing proponents.  The word “exascale” is not mentioned.  “Computing” appears once and “Data” appears eight times.

EOP’s external advisory council on S&T matters is the President’s Council of Advisors on Science and Technology (PCAST).  PCAST is fairly transparent.  Its documents are made public and its meetings are webcast.  Its two most recent documents relevant to the current discussion, Networking and Information Technology R&D (2013) and Networking and Information Technology R&D (2010), emphasize networking and information technology – but not HPC as we know it – and certainly not exascale computing.

Legislative Branch

Just as most of us in HPC don’t have much expertise on matters of policy and law, most members of Congress don’t know much about HPC – or science and technology in general.  To compensate for this lack of expertise, Congress had an Office of Technology Assessment (OTA).  Its mission was to provide members of Congress and Congressional Committees with objective and authoritative analysis of scientific and technical issues.  OTA was eliminated in the mid-1990s.

The House Committee on Science, Space, and Technology is the venue in which many of the HPC discussions on Capitol Hill take place.  For example, here are links to archived webcasts of a couple of relevant hearings that took place earlier this year:

Next Generation Computing and Big Data Analytics

America’s Next Generation Supercomputer: The Exascale Challenge

The S&T policy process in the Legislative Branch is fraught with complexities, short on “objective and authoritative analysis”, and subject to all the partisan strategies and tactics we read about on an almost daily basis in the news.

Judicial Branch

When disagreements about S&T policy cannot be resolved by other means, they may reach the courts.  So, one might ask what objective and authoritative sources are available to them.  The answer is pretty short.  They rely on S&T writings and expert testimony.  Writings and expert opinions can be found to support just about any position.  So, the situation is similar to that in the Legislative Branch.

HPC Policy

HPC policy gets closer to the implementation levels in the federal technology agencies through the Networking and Information Technology Research and Development (NITRD) Program.  NITRD stems from the High-Performance Computing (HPC) Act of 1991 and serves as a venue for agencies to coordinate their HPC program plans and intended expenditures.  NITRD operates under the aegis of the National Science and Technology Council (NSTC).

Among its library of many informative documents, a couple of notable products put out by NITRD are its Strategic Plan and the annual “HPC Blue Book” to accompany the President’s Budget Request to Congress.

The current NITRD Strategic Plan outlines “three essential foundations for advancing leadership in the digital world”: expanded human-computer partnerships; trust and confidence in computing and networking systems; and education of a cyber-capable workforce.  Thus, there is no high-level emphasis on exascale computing or HPC as we know it.

Where the Rubber meets the Road

After all of the processes described above are said and done, the real action on HPC plays out inside the individual technology agencies.  Within the bounds of their agency’s mission and policies, program managers have at least some discretion in carrying out their jobs.  So, if they need advanced computing capabilities to accomplish their missions, they can commit resources to develop or acquire them.  However, since no agency has the mission to advance HPC technologies for their own sake, there are limits to what can be accomplished in this way.

Despite the best efforts of many intelligent and dedicated people within the federal S&T agencies, there are some frustrating limitations that are more or less inherent in the “system”.  Although the usual justifications for federally funded R&D are taking more risk and engaging in longer term research than industry can, agency reward systems may encourage their program managers to minimize risk and think in the shorter term.  The political appointees who run the agencies are only in place for short periods of time, say 4 to 8 years – and more likely the former.  Their main concern is usually creating something new.  The odds run strongly against the same topic being picked by successive appointees.  So, continuity is hard to come by.

In most S&T agencies, a great deal of attention is paid to “community” opinions and advice.  For example, most programs of any size have external advisory committees.  While such community involvement sounds like – and can be – a very positive influence, it often falls short of that goal.  In practice, “community” often means “the collection of people currently known to us”.  Without a lot of effort to keep them vital and open to new membership and ideas, external advisory committees may have fairly limited visions about the larger communities they are seen to represent.

Other external agents interested in influencing program directions include those acting in behalf of laboratory managers, vendors, and those concerned about the long-term health of local economies in places where there are significant federal expenditures.  The advice coming from such sources can be quite valuable, but may be very conservative and narrowly focused.

Most agencies have no policy for rotating personnel.  So, civil servants may stay put in positions for very long tenures. Given human limitations – we only know what and whom we know – this lack of mobility may cause even the best programs to go stale over time.

None of this should come as a surprise. But all of it begs the question: Where are the agents for risk-taking and change? These elements are critical to the timely advancement of rapidly evolving fields like HPC.

More Heat than Light

The HPC policy system appears to be broken. It is based on S&T policies that are decades old and haven’t evolved to keep up with the current state of science and technology, in general, and HPC, in particular.

Although it is widely denied, the Top500 List has filled the void and appears to have become a replacement for HPC policy. An often voiced argument for advancing HPC is that we need to be “number one” in HPC – based, of course, on the Top500 List. This is not an adequate justification for increased HPC expenditures.  There are other important areas in which the US is not number one.  Given limited resources, how shall we prioritize them for action? We need a more compelling HPC story.

Lots of people and organizations are producing lots of documents and holding hearings and other events – but not much seems to be happening.

Action Plan

The absence of a viable HPC policy, compounded with the disarray in the federal budgeting process, has led to a free-for-all.  There are lots of folks advocating lots of things – for example, to list a few of the more popular topics: Big Data; Networking; Information Technology (whatever that is); Education; Industrial Computing; Data Privacy; Open Data; and Exascale Computing.

What we need is some way to reinsert an HPC vision and policy that represent the best interests of our country.  Unfortunately, a coherent and rational HPC vision and policy are not likely to come from the usual channels.  If we accept this to be true, then we need to act outside those usual channels.

First, we need to become well informed about the HPC expenditures currently being made. We need the data and – modulo legitimate privacy considerations – all of the data. We need it now – not at some indeterminate time in the future.

The annual budget process starts with the submission of the President’s Budget request to Congress. This document provides a coherent look at the government’s current and planned HPC expenditures and describes, to a certain level of detail, how those funds are to be spent. This isn’t a perfect starting point for us – but it’s the best one available. Of course, we’ll need the corresponding information about the allocations of federal computing and networking resources.

Let’s take this information and:

  • Build it out so that it drills down to every unclassified HPC project and allocation of computing and networking resources in every relevant agency;
  • Make it openly available in a useful digital form;
  • Provide a basic set of tools to analyze it;
  • Convene a continuing, informed, public discourse; and
  • Reach our own consensus on an HPC vision and policy.

Of course it will take some resources to do this – but we need to step up and make it happen. We need to shine some daylight all the way down into federal HPC and form our own opinions about how to go forward. We may not be able to get all the information we need, but we’ll know what information needs to be broken loose. Our analyses may not be perfect, but they currently have no competition.  We may not reach consensus, but we’ll know where we disagree – and why. If we can accomplish these steps, further steps will either be obvious or unnecessary.

Let’s get engaged.  HPC and S&T policy are simply too important to leave in the hands of the “professionals”.

_____________________________________________________________________________________

Appendix: S&T Policy Players

Executive Branch

Assistant to the President for Science and Technology Policy (APST)

In the Executive Branch, the APST is the highest ranking S&T policy position. It is not a Cabinet-rank position but is close to it, in that it permits direct access to the President. Sometimes this position is filled with the OSTP Director (e.g. current Administration) – and sometimes it’s left vacant (e.g. previous administration). In the event that there is an APST, that person co-chairs the National Science and Technology Council (see below) with the President.

Office of Science and Technology Policy (OSTP)

OSTP is part of the Executive Office of the President (EOP). OSTP’s mission is to serve as a source of S&T analysis and advice for the President with respect to major policies, plans, and programs of the federal government. As mentioned above, the OSTP Director sometimes also serves as the APST.

Office of Management and Budget (OMB)

OMB, also part of the EOP, holds the Executive Branch purse strings.  So, since money fuels S&T, OMB is a very important S&T policy player.  In fact, the OSTP and OMB directors jointly issue an annual memo to federal agencies on science and technology priorities.  The memo on Science and Technology Priorities for the FY 2015 Budget is an interesting read for exascale computing proponents.  The word “exascale” is not mentioned.  “Computing” appears once and “data” appears eight times.

National Science and Technology Council (NSTC)

NSTC is the Executive Branch’s internal coordination body for science and technology issues.  It is responsible for developing and implementing the strategies to carry out the Administration’s S&T agenda.

Operating under the aegis of the NSTC is its HPC policy arm, the Networking and Information Technology Research and Development (NITRD) Program.  NITRD stems from the High-Performance Computing (HPC) Act of 1991 and serves as a venue for agencies to coordinate their HPC program plans and intended expenditures.

Among its library of many informative documents, a couple of notable products put out by NITRD are its Strategic Plan and the annual “HPC Blue Book” to accompany the President’s Budget Request to Congress.

The current NITRD Strategic Plan outlines three essential foundations for advancing leadership in the digital world:

  • WeCompute – Expanded human-computer partnerships, including more capable, available, and affordable systems; more powerful digital tools for people; and new forms of collaboration between the two
  • Trust and Confidence – The ability to design and build systems with levels of security, safety, privacy, reliability, predictability, and dependability that “you can bet your life on”
  • Cyber Capable – Transformed education and training to ensure that current generations benefit fully from cyber capabilities and to inspire a diverse, prepared, and highly productive next-generation workforce of cyber innovators.

President’s Council of Advisors on Science and Technology (PCAST)

PCAST is the external advisory council for the federal S&T enterprise.  It is fairly transparent.  Its documents are made public and its meetings are webcast.  Its two most recent documents relevant to the current discussion, Networking and Information Technology R&D (2013) and Networking and Information Technology R&D (2010), emphasize networking and information technology – but not HPC as we know it – and certainly not exascale computing.

Legislative Branch

Office of Technology Assessment (OTA)

Once upon a time, Congress had an Office of Technology Assessment.  Its mission was to provide members of Congress and Congressional Committees with objective and authoritative analysis of scientific and technical issues.  During its twenty-four-year life it produced about 750 studies on a wide range of topics.  These may be found at the OTA legacy site maintained by Princeton University.  Not feeling the need for OTA’s advice, the 104th Congress withdrew funding for OTA and it disappeared into history.  To some extent, the Government Accountability Office (GAO) has tried to fill the void created by OTA’s disappearance.

House and Senate Committees

The House Committee on Science, Space, and Technology is the venue in which many of the HPC discussions on Capitol Hill take place.  For example, here are links to archived webcasts of a couple of relevant hearings that took place earlier this year:

Next Generation Computing and Big Data Analytics

America’s Next Generation Supercomputer: The Exascale Challenge

In the Senate, a couple of committees to watch for relevant hearings are:

Senate Committee on Energy & Natural Resources

Senate Committee on Commerce, Science, & Transportation

The good news is that such hearings are public and usually webcast.  The less good news is that they are very formal affairs and not really conducive to a thorough exploration of policy alternatives.  This situation is compounded by the absence of the “objective and authoritative analysis” formerly available from OTA.

The committees mentioned above are authorizing committees.  That means that they have jurisdiction for particular topics and can write legislation authorizing the expenditure of specific amounts of federal funds for specific activities related to those topics.  However, the authorization process does not set those funds aside.  The actual expenditures are determined through a separate appropriations process, handled by different committees.  The process is complicated, so suffice it to say that, from an S&T policy perspective, the relationship between authorizing and appropriating is somewhat similar to that between OSTP and OMB in the Executive Branch.  For more about the budget process, take a look here.

So, the S&T policy process in the Legislative Branch is fraught with complexities, short on “objective and authoritative analysis”, and subject to all the partisan strategies and tactics we read about on an almost daily basis in the news.

Judicial Branch

When disagreements about S&T policy cannot be resolved by other means, they may reach the courts.  So, one might ask what objective and authoritative sources are available to them.  The answer is pretty short.  They rely on S&T writings and expert testimony.  Since writings and experts can be found to support either side of most issues, judicial resolution of S&T issues does not inspire optimism.

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