An IBM technology policy expert examines the state of U.S. high performance computing, its role as a driver of innovation and discovery, and the importance of legislation that will sustain America’s leadership in this vital area of technology.
By Mark O’Riley, Technology Policy Executive, IBM Government and Regulatory Affairs
Over the past six decades, the U.S. government, has sponsored the development and deployment of ever more capable high performance computing (HPC) computing systems, which in turn drives remarkable advances in the state of the art of high end computing, and establishes U.S. dominance in the area.
Government funding has played a vital role in seeding the information technology industry. Federal funding also has been instrumental in creating the human resources that drive the innovation process. Without this support, the considerable progress we see today would not have been realized. Through the public, industry, and academic partnership, these efforts have produced ground breaking research in science, manufacturing, aerospace and healthcare, among other industries.
To formally solidify the nation’s support of advanced computing, the High Performance Computing Act of 1991 was signed into law. As then stated in the bill “advances in computer science and technology are vital to the Nation’s prosperity, national and economic security, and scientific advancement.” By 1997, we realized the first teraflop system with ASCI Red at Sandia National Laboratory (equivalent to roughly 10 of today’s laptops).
Again in 2005, as technology advanced and the global competitive landscape evolved, the High Performance Computing Act of 2005 was signed into law providing for “long term basic and applied research on high performance computing.”
Both bipartisan efforts assured the American leadership in high performance computing would not only be maintained but flourish. In 2007, at Los Alamos Laboratory, we realized the first petaflop supercomputer (Roadrunner), a system delivering a peak performance of 1.45 petaflops (roughly 15 thousand high performance laptops).
Our capabilities have not gone unnoticed in the global community where today leadership high capability systems can be found in countries far and wide including China, Japan, the United Kingdom, and Germany amongst others. In 2014, the Department of Energy awarded three contracts, one each to Oak Ridge National Laboratory, Argonne National Laboratory, and Livermore National Laboratory to develop platforms with at least 100 times the capability of Roadrunner. In the space of seven years, our national research and development programs have further cemented the United States position.
Now the time has come again to revisit the nation’s policy of High Performance Computing. In both the House of Representatives and the Senate, bills have been introduced updating the 2005 Act to reflect the ever increasing pace of technological development. H.R. 874 and S. 454, both bipartisan efforts, would position the nation for the development of an exascale platform, which would produce a peak performance improvement of at least 700 times over the 2007 platform. The capabilities of this system should include the ability to manage the explosion of Big Data volumes we expect to see in the coming years, as well as to perform complex analytics and gain important insight from that data, allowing research in many fields to accelerate at a pace that we could only imagine some 60 years ago.
This enormous increase in computational power, tied to the platform’s data centric capabilities, would enable discovery to take very significant steps forward.
We applaud both the Congress and the Administration for supporting this effort because, as we opened, “if nothing changes, nothing changes.”