Survey: There is No Exascale Roadmap

Print Friendly, PDF & Email

In a recent survey by The Exascale Report, 63 percent of HPC community respondents said they do not believe there is a roadmap that will get us to exascale.

Due to a failure of vision and courage shared across the necessary constituent institutions, both political and technical, there is no consensus and supported roadmap, in the US or elsewhere that will deliver the necessary innovation and persevere through the daunting challenges to the final realizable conclusion,” writes Thomas Sterling from the Indiana University. “But there could be; there should be.”

In addition to Sterling, the Exascale Report Community Response Summary includes comments from a veritable who’s who of HPC thought leaders including Dona Crawford (LLNL), Andrew Jones (NAG), Kimmo Koski (CSC), and Wilf Pinfold (Intel).

Comments

  1. Exa = 1000 petaflop/s, peta= 1000 tera/s giga = 1000 megaflop/s
    Around 1940, WWII, Morse telegram, computer size of a basketball court able to do at least couple Megaflop/s.
    2000-2001; the world achieves a laptop – 1/5 size of a suitcase, able to do 600mhz – 1.9 ghz
    (pentium I – Intel duo (first gen.)
    2001 – 2008: Four cores, six cores, eight cores, 12 cores, 16 cores, (Intel even tested 100 cores; or some private enterprise had released 6.7 ghz die).
    2008 – 2012: We hit E357, Amd Opteron 8c, 12c, 16c… APUs, Intel’s I3,5,7… At least 4 Ghz into a 1/9 size of a suitcase laptop.
    >>>
    Point being it is 52 years since 1940. Petaflop computer is possible, proven by the American Roadrunner, and continued by future people who dedicate their efforts to improve present computer architects to achieve the “petaflop” standard. In another 20 or 30 years, the Exaflop is able to be available; and possibly petaflop supercomputers are commerially available to enterprises…
    In another 100 years, I think humans are able to beat cancer with technology. Technology can give humanity an edge on bioinfomatics, biogenetics, hybrid platforming on creating nanoscaled androids, and more.