The Future of Supercomputing at NASA
Total Page:16
File Type:pdf, Size:1020Kb
Southwestern Oklahoma State University SWOSU Digital Commons Student Research Business & Computer Science 3-8-2019 The Future of Supercomputing at NASA Ezgi Gursel Southwestern Oklahoma State University Jeremy Evert Southwestern Oklahoma State University Follow this and additional works at: https://dc.swosu.edu/cpgs_edsbt_bcs_student Recommended Citation Gursel, Ezgi and Evert, Jeremy, "The Future of Supercomputing at NASA" (2019). Student Research. 16. https://dc.swosu.edu/cpgs_edsbt_bcs_student/16 This Poster is brought to you for free and open access by the Business & Computer Science at SWOSU Digital Commons. It has been accepted for inclusion in Student Research by an authorized administrator of SWOSU Digital Commons. An ADA compliant document is available upon request. For more information, please contact [email protected]. COLLEGE OF & PROFESSIONAL GRADUATE STUDIES This material is based upon work supported by Dr. Tsengdar Lee, Program Manager of the High• the National Aeronautics and Space Dr. Tsengdar Dr. Stephen Wheat Dr. Henry Neeman End Computing Division, NASA Administration under Grant No. NNX15AK02H NASA Oklahoma Space Grant Consortium Lee Abstract Dr. Tsengdar Lee is the Program Computers are a growing part of everyday life Q1: Why are "Supercomputers are "Supercomputers allow us to model, "Supercomputers are used to solve Manager of the High-End supercomputers considered a scientific simulate, explore, and analyze some of the hardest problems, Computing Program at NASA. in many ways. For some of the biggest and He is responsible for maintaining most interesting problems, people have used important? How instrument. It is important in phenomena at much greater level of scale especially the hardest science and important is the development of new at finer levels of fidelity than can otherwise engineering problems... problems that the HEC capability to some of the biggest and most interesting supercomputing theories (constructing be made possible with single simply won't fit on a regular computer. support the agency's aeronautics computers. This is the general area of High in current and models) and verify/validate systems. Supercomputers are valuable in Here's how to think of it: research, human exploration, Performance Computing (HPC) or future NASA the theory by comparing to getting things done that otherwise could Computational Science & Engineering scientific discovery, and space Supercomputing. One organization with a operations? observations. not be done. makes the impossible possible, and operations missions. Lee is long and storied history with supercomputing supercomputing makes the impossible currently the NASA Weather Focus Area Lead.4 is the National Aeronautics and Space Not only supercomputing will Everything, and I mean everything!!, is practical. Administration (NASA). The supercomputers be used for constructing the modeled. These models are sufficient to theories but also for multi• replace all but a small part of For decades, NASA has had some of at NASA have a variety of missions including dimensional data analysis. experimentation and prototyping, thus the fastest supercomputers in the weather forecasting to helping astronauts at design of space vehicle and enabling work to be done that otherwise world. Like any large science and Dr. Stephen Wheat, Professor of Mathematics the International Space Station. As problems airplane, Earth and space could not be afforded." engineering organization, they're _and_Com uter Science,_Oral Roberts_Universit _ continue to grow complex, the growth of model and theory hugely dependent on supercomputing, supercomputing seems inevitable. development and large for research, design, visualization, and scale data analytics. It's also many other needs." Dr. Wheat has spend 38 years in What are supercomputers? being used to replace the High-Performance simulators of space Computing (HPC) industry. Dr. As a general rule, a supercomputer is environments." Wheat is the recipient of a Gordon something substantially more powerful than - - Bell Prize (1994) as well as the --" Uncertain microprocessor "The largest limitation is always "Budget. The supercomputing what you might have on your desk today-at What, in your Intel Achievement Award (1997), opinion, is the architecture is becoming the budget. The scope of the missions community knows how to make bigger both associated with his advancing least a hundred times more powerful. The biggest obstacle biggest obstacle. NASA's increase; the need for SC increases; etc. supercomputers than we can afford." the technology in HPC. A major supercomputers of today can perform for NASA legacy scientific and But the budget does not HPC publication twice recognized trillions of calculations in a manner of supercomputing technical applications will increase. Additionally, skilled scientists in him as a "Person to Watch". seconds. NASA's biggest supercomputer currently? need to be refactored to the the art of computational science. The He was recognized as one of Pleiades has 935 TB total memory and can new architecture. It would be pipeline of training such people is still too NASA AMES 25 most influential very time consuming and fragile." do up to 7.24 petaflops (101115) people in their 25-year history for manpower intensive. The his contributions to NASA's return computations in a second.1 uncertainty would create to flight post the Columbia space Supercomputers are everywhere in our daily significant schedule delay." shuttle disaster.5 lives. A modern iPhone has more computing What is one thing "Can we improve our Earth "Just about anything, from biology, to "It's a very broad spectrum, and now power than a// of NASA during the Apollo NASA expects to observations so that we can chemistry, to physics, to behavior, and that Al/ML/Deep Learning has been Dr. Henry Neeman, Founding Director for OU Mission in 1969. In other words, your phone learn from its extend the limit of weather well beyond." added into the mix, more than ever." Supercomputing Center has the power to send a man on the supercomputers? forecast to many weeks (up to a season)?" Henry Neeman is the founding moon.2 What Is the "I believe it is the "NASA is already well-set for doing the "In the near term, the biggest recent Director of the biggest development of the high work it needs to do. Computers aren't change has been the incorporation of OU Supercomputing Center for advancement in bandwidth memory which going to get appreciably faster, so the more and more Al/MUDL capability Education & Research the field of allows more and faster data question is whether you can do more with into both CPUs and GPUs. In the (OSCER), Associate Professor of supercomputers? flow from the memory to the more computers." (S. Wheat, personal longer term, as the current CMOS Engineering, and Adjunct processor." (T. Lee, personal communication, February 11, 2019) technology for making chips comes to Associate Professor of Computer communication, February the end of our ability to keep shrinking Science at OU. He also leads the 11, 2019) it, new computing technologies are Virtual Residency, which teaches needed" (H. Neeman, personal research computing facilitation communication, February 17, 2019, so far to over 400 research 2019) computing professionals. He received his BS in Computer Science and his BA in Statistics with a minor in Mathematics in 1987 from the Figure r-Cornparison in processing power and References 1. Naoonal Aeronau!Jcaand SpaoeAdrrin.s.n.1m.(20 17, Oeoerrbef 12). P'9tad8sconflsµ<t.oo 5. R;.;oghom, J. V. (n.d.). Dr. S1ophon R. Who.it F;)C!Aty P'ro6kt m ORU. RetritYod fotwuury 27, University at Buffalo, MS in CS from speed between the Cray-2 supercomputer from decal\S. HECX:: kn<»t.•edgebase. Retneveci February 27, 2019, rrom 2019, from h11p:rJW--.w.oru.oouracaoemlcsllacultyi)l'ol 8&'Protbe.pttp?lc:t-275 htrp;:.O\w.w..n<1J.nm,qovJhoo:)~\>h:iadc$·conflg1.W<1~·dc1;ils_n.hln'll University of Illinois at Urbana• 1985 and the iPhone 4 released in 2010 3 S. Neoman, H. t2017), Herty No~"1\. R.Qcnoved FebnRry'ZT, 201~ from 2. C:sni17C , S. (3'.109. July 13). Nu;i · do-it·yotu:<df podcout:rod!.et ~ution. Aelrit:vt!dFebN:lry h11pi1hneeman.o!)Ce1.ou.eour Champaign (UIUC) in 1990, and his 27, 2019, trom htrpe1/wP.w.oasa.gcwteudteoo&110f'educatorsfd.ypodcasf/n>eket·&'IOllUoo·IOdex· diy.t'M'll 7. SoulhweetemO:lahoma StateU~ver&t)'. (n.d.).Pa1ey PaO<er. Relfll8'Vedf rom PhD in CS from UIUC in 1996. hllp$'itAWW.&WOSl.a.ilXU'.leadomlcs41}1bio:;fpa:l5)'•p.1rkGr.a:;px 3. FT!po 1· ~rt: Exchange. (n..d). Pr0Qfl$sing Power Comp.:uod[DigMI imotgeJ. Retrieved Fobtu;:uy 25. 2019, ~ http;:l1'p;.gos.ollf)Cf15~1Cch;,.ngo.co~rcN>OWOf-C::Qnl)lltCd 8. Soulhw°'tcm G:.~ Stitc-Uriversity. (2018. ~by 23). Leri G'Ayn '1J'T'ICd Diroc:tor of $pQn50fedP'rwriJTI& al $\\IOSU. Retri9vtd FGbru;,.ry 27, 2019, from 4. N:Hion:tl Aet~udo: and S~ Adrrimtra6on. (n.d.) NASA ~·ond ccmp.iting: pn>gram hllpS:lh.ww.SW09u.edu.rt~W..Vr8'8aaeil201&'2018-0S..23a.asp:t maneaer. R001eved Fel>Net)' 11, 2019, lromttU>S:IMww.nec..nasa.QOYJ'aboutl\eeJ'l'lll'f .