DOC PREVIEW
UW CSEP 590 - A History of Supercomputing

This preview shows page 1-2-3-20-21-40-41-42 out of 42 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 42 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 42 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 42 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 42 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 42 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 42 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 42 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 42 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 42 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

A History of SupercomputingOutlineDefinitionsNames and PlacesEarly days: ColossusColossusEarly days: EniacEniacOther early supercomputersMainframesMainframes: LARCMainframes: Stretch and HarvestMainframes: CDC 6600CDC 6600 ConsoleMainframes: IBM 360/91IBM 360/91 ConsoleMainframes: CDC 7600Two CDC 7600sSIMD arrays: Illiac IVILLIAC IVSIMD Arrays: PEPE and BSPOne-bit Arrays: STARAN, DAP, MPPOne-bit Arrays: Connection MachineVector pipelining: TI ASCVector Pipelining: CDC Star 100Amdahl’s lawVector Pipelining: Cray-1Cray-1VLIW: FPS, Multiflow, and CydromeShared Memory: Cray Vector SystemsShared Memory: MultithreadingShared Memory: CC-NUMADistributed Memory: Cosmic CubeDistributed Memory: Classic MPPsDistributed Memory: ClustersThe Top 500 ListJapanese Systems: Fujitsu, NEC, HitachiJapanese National ProjectsDARPA Strategic Computing ProgramThe DOE ASCI ProgramWhat Next?Apologies: Companies Not Mentioned1A History of SupercomputingBurton SmithMicrosoft2OutlineDefinitionsEarly days: Colossus and EniacMainframes: Univac LARC to CDC 7600SIMD systemsProcessor arrays: Illiac IV, PEPE, and BSPOne-bit arrays: Staran, DAP, and the Goodyear MPPVector pipelining: TI-ASC, CDC Star-100, and Cray-1 Wide instructions: FPS, Multiflow, and CydromeMIMD systemsShared memory: Cray X-MP to SGI AltixDistributed memory: Cosmic Cube to Blue GeneJapan-US competitionWhat next?3DefinitionsSupercomputer: the world’s fastest computer at the timeSIMD: Single Instruction Stream, Multiple Data StreamOne instruction at a time, each operating on an array of dataMIMD: Multiple Instruction Stream, Multiple Data StreamMultiple processors asynchronously issuing instructionsShared Memory: MIMD computer in which a common memory is accessible by all processorsUMA: Uniform memory access by all processorsNUMA: Non-uniform memory access, based on placementDistributed Memory: MIMD computer in which the memory is partitioned into processor-private regionsRISC: Reduced Instruction Set ComputerUses fast and simple instructions to do complex thingsPipelining: Processing in assembly-line style4Names and PlacesBMD/ATC: Ballistic Missile Defense Advanced Technology Center, Huntsville, ALBRL: Ballistics (now Army} Research Laboratory, Aberdeen, MDDARPA: Defense Advanced Research Projects Agency, VANASA-Ames: Ames Research Center, Mountain View, CANASA-Goddard: Goddard Space Flight Center, Greenbelt, MDLivermore: Lawrence Livermore National Laboratory, Livermore, CALos Alamos: Los Alamos National Laboratory, Los Alamos, NMNSA: National Security Agency, Fort Meade, MDSandia: Sandia National Laboratories, Albuquerque, NM5Early days: ColossusUsed at Bletchley Park, England during WW IIfor cryptanalysis of Lorenz SZ40/42 rotor systems (“Fish”)Only recently declassified by Her Majesty’s GovernmentConcepts by Max Newman et al.Designed and built by Tommy Flowersat the Post Office Research Station, Dollis HillFeatures:Paper tape loop data input at 30 mph = 5000 characters/sec1500 vacuum tubes (“valves”) in the Mark I, 2400 in Mark IIProgrammed by switchesNot a general-purpose computer6Colossus7Early days: EniacBuilt at Penn’s Moore School for BRL, Aberdeen, MDThe motivating application was artillery firing tablesIts first use was evaluating H-bomb feasibilityDesigned and built by J. Presper Eckert and John MauchlyUnveiled at the Moore School on February 15, 1946It had 17,468 vacuum tubes and consumed 174 KWReliability was a concern, so it was left on continuouslyIt was 80 feet long and weighed 30 tonsProgramming was via switches and patch cablesNot a general-purpose computer8Eniac9Other early supercomputersZuse Z3 (1941)The IAS machines (1952)Manchester/Ferranti Mark I (1951)Univac 1 (1951)10MainframesThe 50’s and 60’s saw fast progress in computer technologyMost of it was first used in “mainframe” supercomputersThese systems were used for both business and scienceLater, supercomputers became much more science-orientedExamples of new technology included:Magnetic core memoryTransistor logic circuitsFloating point hardwarePipeliningHigh level languages and compilers11Mainframes: LARCBegun in 1955 for Livermore and delivered in 1960Had dual processors and decimal arithmeticEmployed surface-barrier transistors and core memory12Mainframes: Stretch and Harvest• IBM 7030 (STRETCH)• Delivered to Los Alamos 4/61• Pioneered in both architecture and implementation at IBM• IBM 7950 (HARVEST)• Delivered to NSA 2/62• Was STRETCH + 4 boxes• IBM 7951 Stream unit• IBM 7952 Core storage• IBM 7955 Tape unit• IBM 7959 I/O Exchange13Mainframes: CDC 6600Seymour Cray and Bill Norris left Univac to start CDCCray’s CDC 1604 was the first successful transistor systemHe wanted to build a scientific supercomputerThe 6600, built with silicon transistors, shipped Sept. 1964The chief architects were Seymour Cray and Jim ThorntonThe 6600 had several notable featuresA very simple instruction set, lending itself to more speedNo decimal arithmetic needed or wanted10 arithmetic function units able to process in parallelThis overlapped execution style is now called superscalarIt was coordinated by a famous circuit called the scoreboardParallel I/O processors using a new idea called multithreadingSome say RISC means “Really Invented by Seymour Cray”A futuristic operator’s console was provided at no extra charge14CDC 6600 Console15Mainframes: IBM 360/91IBM’s Tom Watson was angry: "Last week, Control Data ...announced the 6600 system. I understand that in the laboratory...there are only 34 people including the janitor. Of these, 14 are engineers and 4 are programmers. Contrasting this modest effort with our vast development activities, I fail to understand why we have lost our industry leadership position by letting someone else offer the world's most powerful computer." The 360/91 (AKA 370/195) was IBM’s answerIt was delivered to NASA-Goddard in October 1967Killer App: Passenger Airline Reservation Systems (1964)Jointly developed by IBM, American, and United AirlinesOriginally written in IBM assembly (i.e. machine) languageIt was very close to the CDC 7600 in performanceThe 360/91 also had a pretty impressive console…16IBM 360/91 Console17Mainframes: CDC


View Full Document

UW CSEP 590 - A History of Supercomputing

Documents in this Course
Sequitur

Sequitur

56 pages

Sequitur

Sequitur

56 pages

Protocols

Protocols

106 pages

Spyware

Spyware

31 pages

Sequitur

Sequitur

10 pages

Load more
Download A History of Supercomputing
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view A History of Supercomputing and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view A History of Supercomputing 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?