Presentation is loading. Please wait.

Presentation is loading. Please wait.

2015/5/31 Fifty Years of Japanese HPC ─From transistor to Exascale─ Yoshio Oyanagi Education Center for Computational Sciences Kobe University Fifty years.

Similar presentations


Presentation on theme: "2015/5/31 Fifty Years of Japanese HPC ─From transistor to Exascale─ Yoshio Oyanagi Education Center for Computational Sciences Kobe University Fifty years."— Presentation transcript:

1 2015/5/31 Fifty Years of Japanese HPC ─From transistor to Exascale─ Yoshio Oyanagi Education Center for Computational Sciences Kobe University Fifty years of Japanese HPC

2 Outline of my talk  Historical overview of HPC in Japan, as contrasted to US, Europe, China, …..  What is the difference between Japan and US trends? –In Japan, big electronic companies manufactured vector and parallel supercomputers besides main frames. –In US, venture companies were active in vector (Cray, Convex etc.) as well as parallel (TMC, FPS, Meiko, nCUBE, KSR, ….) machines. 2015/5/32Fifty years of Japanese HPC

3 “50 Years of HPC” A Column in HPCwire Japan (Since July, Updated every Monday.) 2015/5/33 (Sorry in Japanese) Fifty years of Japanese HPC

4 1960’s in Japan  OKITAC-5090 (1961)  Transistors + Diodes + core memory  Floating point (decimal 10+2) –Add/sub 0.4 ms –Mult 4.9 ms (peak 0.4 kFlops) 2015/5/34 Not very high performance !! Fifty years of Japanese HPC

5 1960’s in US  CDC6600 IBM System/360 –4 Mflops (peak)  Contract with DARPA for the ILLIAC IV 2015/5/35 Pictures from Wikipedia Fifty years of Japanese HPC

6 1960’s Japan USA  FACOM 270  HITAC 5020, 8000  TOSBAC 3400  NEAC 2200  MELCOM 3100, 9100  OUK 9400  FACOM  IBM 360 model 91 (1967)  IBM 360 model 85 (1968)  Intel founded (1968)  IBM 2938 array proc. (1969)  CDC7600 (1969)  AMD founded (1969) 2015/5/36Fifty years of Japanese HPC

7 2015/5/37 Primordial Ages(1970 ’ s) Japan USA/Europe 1970 IBM System/ CDC STAR-100 announced 1972 Goodyear STARAN 1974 DAP, BSP and HEP started 1975 ILLIAC IV operational 1976 Cray-1 shipped to LLNL 1976 FPS AP-120B 1977 Siemens SMS HEP single processor operational (red: vector architecture) 1970 FACOM FACOM APU 1978 HITAC M-180 IAP 1978 PAX project started 1979 HITAC M-200H IAP 1979 MELCOM COSMOIII IAP 1981 MITI Supercomputer Project started( ~ 89) (green: parallel architecture) Fifty years of Japanese HPC

8 2015/5/38 Characteristics of Japanese vectors 1. Japanese vector processors only a couple of years after Cray Manufactured by main-frame vendors with semiconductor facilities (not ventures) Vector processors are attached to mainframes 3. FACOM APU (first Japanese vector) a) CPU and APU share the main memory (1 MW×36bits) Heterogeneous architecture b) Vector register of 1792 words c) list vector and DO-loop with IF supported d) 22 Mflops (single), 11 Mflops (double) e) installed in NAL in August 1977 Fifty years of Japanese HPC

9 Characteristics of Japanese vectors 4. HITAC IAP (1978-) a) memory-to-memory (no vector registers) b) summation, inner product and 1st order recurrence could be vectorized in FORTRAN c) vectorization of loops with IF ’ s (M280) d) M-200H IAP (1979): 48 Mflops, M-280H IAP (1982): 67 MFlops 5. NEC ACOS 1000 IAP (1982) a) 36 bit machine b) peak 28 MFlops 2015/5/39Fifty years of Japanese HPC

10 2015/5/310 1st Generation (1H of 1980 ’ s) Japan US/others 1981 CDC Cyber Cosmic Cube Project 1982 Cray X-MP/2 (420 MF) 1982 Alliant FX/8 delivered 1982 HEP installed 1983 Encore, Sequent and TMC founded, 1983 ETA span off from CDC 1984 Cray X-MP/4 (820 MF) 1984 Legend group in China started 1985 Convex C Intel iPSC/1, T414, nCUBE/1, Stellar, Ardent 1985 Cray-2 (1.952 GF) 1986 CM-1 shipped 1986 FPS T-series 1980 PAX MITI Supercomputer Project ( ~ 89) 1982 Fifth Generation Project ( ~ 92) 1982 NEC ACOS-1000 IAP 1982 HITAC M280H IAP 1983 HITAC S-810/20 (630 MF) 1983 FACOM VP PAX-64J 1985 FACOM VP-400 (1.142 GF) 1985 NEC SX-2 (1.3 GF) Fifty years of Japanese HPC

11 2015/5/311 Characteristics of Japanese SC in the 1st Generation 1. Compatibility with the mainframes 2. Single processor with multiple pipes 3. Large main memory (256MB vs. X-MP 32MB) 4. Large vector registers 5. List-directed vector instructions (gather/scat) 6. Different control of vector units 7. No commercial parallel machines Fifty years of Japanese HPC

12 Characteristics of Japanese SC ’ s in the 1st Generation 8. Aggressive installation of SC’s in universities and laboratories –7 SC centers open to all university people –Laboratory SC’s for physics, chemistry, …  In US, most SC’s are in DOE, DOD, NASA, not available for university researchers  Lax Report (1982) –5 NSF centers in /5/312Fifty years of Japanese HPC

13 2015/5/313 Characteristics of Japanese SC ’ s in the 1st Generation 9. MITI Supercomputer Project ( fy)  Started before the Japanese supers (S810, VP, SX)  Targets: –10 GFlops parallel vector machine (PHI), –Dataflow machine (sigma-1) –Josephson device, MPP etc.  In alliance with six companies Fifty years of Japanese HPC

14 2015/5/314 2nd Generation (2H of 1980 ’ s) Japan US and Europe etc. 1986“863” Plan in China 1986Manheim Supercomputer seminar (→ISC) 1987 ETA-10 (10 GF) 1987CM Cray Y-MP (2.66 GF) 1988Intel ipsc/2 1988First Supercomputing Conference in Orlando 1989ETA shut down, JvN SC shut down 1989BBN TC2000, Myrias SPS-2, Meiko CS, NCube2 1990Intel ipsc/860, MasPar MP HITAC S-820 (3 GF) 1989 FACOM VP2600 (5 GF) 1990 MITI Supercomputer Project ended 1990 NEC SX-3 (22 GF) 1990 QCDPAX completed Fifty years of Japanese HPC

15 2015/5/315 Characteristics of Japanese SC in the 2nd Generation 1. Vector multiprocessor appeared in Japan with modest multiplicity 2-4 vs. 8 (Y-MP, ETA) 2. The semiconductor technology developed for the vector is transferred to mainframes (Used to be: mainframe→vector) 3. Still no commercial MPP ’ s in Japan, although parallel research was active in academia Fifty years of Japanese HPC

16 Installation of SC’s in Japan (1987-8)  Cray Research (Cray-1, X-MP):10  ETA (ETA-10): 1  Hitachi (S810):15  NEC (SX-1/2):7  Fujitsu (VP-50, 100, 200, 400): /5/316Fifty years of Japanese HPC

17 US-Japan Trade Conflicts  1985/9: Plaza Accord (G5)  1985: SX-2 was cancelled by NCAR after bidding  1986/9: US-Japan Semiconductor Agreement  1987: Japanese SC was cancelled by MIT after bidding 2015/5/317Fifty years of Japanese HPC

18 2015/5/318 Super 301 Act  In 1989, US government decided to apply “ Super 301 Act ” (Omnibus Trade and Competitiveness Act of 1988) to Japan and identified three items including supercomputers.  Washington threatened Japanese governmental institutions to purchase US supercomputers. (Titech ETA10, ETL X-MP)  On the other hand, US governmental institutions had no Japanese supercomputers at all. Fifty years of Japanese HPC

19 2015/5/319 3rd Generation (1H of 1990 ’ s) Japan USA/Europe 1990 MasPar MP Intel iPSC/ HPCC started (-96) 1991 Cray Y-MP C90 (16 GF) 1991 Intel Paragon, TMC CM FPS bankrupt 1992 MasPar MP Top500 started 1993 HPCN started ( ~ 2001) 1993 Cray T3D, CS6400, nCUBE2, KSR SSI shut down 1993 IBM SP Cray SP TMC, KSR Chapter RWCP started 1992 CP-PACS started 1993 Fujitsu NWT 1993 Fujitsu VPP HITAC S-3800 (32 GF) 1993 NEC Cenju Fujitsu AP NEC SX-4 Fifty years of Japanese HPC

20 2015/5/320 Characteristics of Japanese SC in the 3rd Generation Drastic changes in Japanese vectors 1. Hitachi (S3800) Shared memory vector parallel processor using ECL technology 2. Fujitsu (VPP500) Distributed memory vector parallel proc. using GaAs as well as silicon tech. 3. NEC (SX-3) Cluster of shared memory vector parallel Unix as host OS 4. No action to “ Attack of killer micros.”(SC90) ◆ Commercial MPPs, sold as a parallel testbed Fifty years of Japanese HPC

21 Trends in US (1 st half of 1990’s)  Active vector machines: Cray Y-MP, C90 and Convex C2  MPP:  Performance of COTS CPU’s was increasing drastically  custom CPU → commodity CPU chips  ventures company → big companies (IBM, Cray, Intel)  Strong national initiative: HPCC, NII, HPC Act /5/321Fifty years of Japanese HPC

22 HPCC in US  Blue Book (1991/2, G.W.Bush) – “Grand Challenges: High Performance Computing and Communications”  The High Performance Computing Act of 1991  Trade conflicts: –“Japan should buy more US supers” (M. Kantor, 1993/4) Otherwise, possible retaliation. –6 out of 11 in public sector bought US machines in the 1994 supplementary budget.  In Europe, Rubbia Report (1991/1) –European Teraflop Initiative 2015/5/322Fifty years of Japanese HPC

23 2015/5/323 4th Generation (2H of 1990 ’ s) Japan USA/Europe/China 1995 DOE: ASCI started 1995 Cray T90 (57.6 GF) 1995 Cray Computer bankrupt 1996 Dawning in China started 1996 Cray T3E 1996 Cray Research merged into SGI 1996/10 Serymour died 1996 nCUBE merged into Oracle 1996 MasPar went out of HPC 1997 ASCI Red (Intel) 1997 NSF: NPACI started 1998 Cray SV ASCI Blue Pacific (IBM) 1998 ASCI Blue Mountain (SGI) 1995 Fujitsu VPP cp-pacs completed 1996 Hitachi SR Fujitsu VPP Fujitsu AP NEC Cenju Hitachi SR NEC SX-5 Fifty years of Japanese HPC

24 2015/5/324 Characteristics of Japanese SC in the 4th G. 1. Fujitsu and NEC seem to follow the line of conventional vector supercomputers in CMOS Fujitsu: distributed memory NEC: cluster of shared memory nodes 2. Hitachi adopted RISC (pseudo)vector architecture. SR2201: sliding window(proposed by Tsukuba) SR8000: 8 tightly coupled CPUs in one node (self-developed CPU with Power architecture) Fifty years of Japanese HPC

25 US Trends  ASCI Project (originally ) –SC’s in LANL, LLNL, SNL –Red (1+), Blue (3+), White (10+), Q(30+), Purple (100+) --- every 2 years  PITAC (I: , II: ) –IT2 project  NPACI ( )  Petaflops I in Pasadena (1994) Petaflops II in Santa Barbara (1999) 2015/5/325Fifty years of Japanese HPC

26 NCAR Procurement  1996/5/20 UCAR selected SX-4 for NCAR –finalists were C90, SX-4, VPP700  Criticism by Cray and Obey –Tax from US people should be used for US competitiveness –NEC (and Fujitsu) was dumping  1996/7/29 Cray filed an apeal to DoC and ITC for dumping  1997 –454% anti-dumping customs to NEC supercomputers –388.7% to Fujitsu 2015/5/326Fifty years of Japanese HPC

27 2015/5/327Fifty years of Japanese HPC Japanese machines in Top 20

28 2015/5/328 Fifth Generation (1H of 2000) Japan US/Europe 2000: Tera became Cray 2000: SGI: Origin : ASCI White (LLNL, IBM) 12 TF 2001: NSF: TeraGrid stared 2002: ASCI Q (LANL, HP) 20TF 2002: DOD: HPCS started 2002: Cray (NEC) SX-6 to ARSC 2004: NASA Columbia (SGI) 64TF 2004: BlueGene/L at IBM 90TF 2004: “The Path to Extreme Computing” conf. (to Exa) 2005: BlueGene/L (360TF) 2005: ASCI Red Storm (SNL) Cray XT3/XT4 2001: NEC: anti-dumping customs cleared 2001: Fujitsu PRIMEPOWER : NEC SX : Earth Simulator 40TF 2002: Fujitsu PRIMEPOWER HPC : NEC SX : NEC SX : Hitachi SR11000 Fifty years of Japanese HPC

29 Japan Trends  NEC kept making vector machines after the Earth Simulator  Fujitsu: PRIMEPOWER (Sparc) and PRIMERGFY (x86)  Hitachi: OEM of IBM servers. SR11000 (Power4+, 6.8 GF/CPU)  NAREGI (2003-8): Grid project 2015/5/329Fifty years of Japanese HPC

30 Japan Trends  IT Strategic Headquarter (2001): e-Japan, but only network was emphasized. –At this stage, level up of supercomputers was to be promoted according to the needs of each field (not a national project).  Earth Simulator attained 36 Tflops (2002)  Information Science and Technology committee in Mext has been discussing the measures to promote computational science and technology since August Fifty years of Japanese HPC /5/3

31 Japanese machines in top /5/3Fifty years of Japanese HPC31

32 US Trends  DOE: ASCI continues –BlueGene and Red Storm added  NSF: Teragrid ( )  DOD: HPCS (High Productivity Computing Systems) –2002: Phase I (IBM, Cray, Sun, HP, SGI) –2003: Phase II (Cray, IBM, Sun) –2006: Phase III (Cray: XC, IBM: PERCS)  HEC Revitalization Act of 2004 and 2005  Exa started in “The Path to Extreme Computing” in Santa Fe (2004/10) 2015/5/332Fifty years of Japanese HPC

33 Sixth Generation (2H of 2000 and later) Japan US/Europe/China  : K Computer P. –2011: 10 PF attained  2006: T2K open SC –2008: installed in 3 univ.  2006: TSUBAME 1.0 –2010: TSUBAME 2.0  2009: Fujitsu fx-1 to JAXA  Cray XT/XE/XC  2005: ASCI Purple (LLNL, IBM)  2006: Rangers (TACC)  2007: NCSA, BlueWaters started (IBM→2011 Cray)  2007: Kraken (Tennessee)  2008: Roadrunner (LANL)  2010: Tianhe-1A (NUDT)  2010: PRACE started  2011: NSF, SXEDE  2013 America COMPETES Act 2015/5/333Fifty years of Japanese HPC

34 History of the K Computer  Recommendation to a Mext committee (2005): –Promote a national project to construct a leading edge supercomputer –Government decision (July 25, 2005)  Riken started the project (October 2005)  Mext funded four projects to promote “Element Technologies for Future Supercomputers” in $40M per year (in total) –Four groups were accepted 1.System Interconnect (Kyushu U and Fujitsu) 2.Interconnect by IP (U of Tokyo, Keio U etc) 3.Low Power Device and Circuits (Hitachi, U of Tokyo, U of Tsukuba) 4.Optical Connection of CPU and Memory (NEC and Titech) Fifty years of Japanese HPC /5/3

35 History of the K Computer  2005: Proposed architectures and killer applications  2007: Hybrid machine with vector and scalar  2009: Five strategic application fields defined  2009: Withdrawal of vector machine  2009: the Government Revitalization Unit proposed to shutdown the Next Generation Supercomputer Project (Nov. 13, 2009)  2009: HPCI started (High Performance Computing Infrastructure, alliance of supercomputer centers and users) Fifty years of Japanese HPC /5/3

36 Original Proposal Large scale processing Scalar computer Special-purpose computer Fifty years of Japanese HPC /5/3

37 Fifty years of Japanese HPC Friday, November 13, 2009 The 3 rd WG of the Government Revitalization Unit “Why should it be No.1 in the world?” “Is No.2 not enough?” Vote : Abolish 1 Postpone 6 Budget shrink 5 Conclusion Freeze the project! ← 村田 ( 謝 ) 蓮舫 (Hsieh Lien Fang) Picture from /5/3

38 No.1 in the world  Pflops attained using 80% of the system ー No.1 in the Top500 (ISC2011) June 20, 2011  Pflops in SC2011 (Seattle)  Open to users: September 2012 –1 st period: 2012/9 to 2014/3 –2 nd period: 2014/4 to 2015/3  RIST is to manage K Computer users (Research Organization for Information Science & Technology) Fifty years of Japanese HPC /5/3

39 What’s next? Toward EXA FLOPS!!  Preliminary consideration among scientists (In US, since 2004)  Mext started a WG for future HPC (April 2011) –Hardware-System Software-Application co-design is important. –Should be science-driven. We identified possible break throughts in science and technology. Social needs are also considered. Linpack Exaflops is not our target. –Limitation by budget, power, …. –Several different architectures are considered. Fifty years of Japanese HPC /5/3

40 Architecture Algorithm Application Fifty years of Japanese HPC /5/3

41 Two subgroups worked  Application Subgroup –Application –Numerical library –Algorithm –Automatic tuning  System Subgroup –CPU and architecture –Compiler –System software Fifty years of Japanese HPC41 Final Report in March /5/3

42 Memory and memory bandwidth Big technical issue Fifty years of Japanese HPC42 Large B/FlopsSmall B/Flops Large B/Flop Small B/Flop 0.1 standard SoC K GPU Small system ES 2015/5/3

43 WG for Future HPCI Policy  2012/4 ~ 2014/3 (25 meetings)  26 members, Chair: Oyanagi  Recommendations –Hierarchical deployment of supercomputers –Leading machines: Flagship system and its complementary systems –Flagship: “exascale” machine toward 2020 –Support to users, especially industry users 2015/5/3Fifty years of Japanese HPC43

44 Flagship system Complementary systems to the flagship Universities and laboratories Leading systems Allocated thru HPCI 2015/5/3Fifty years of Japanese HPC44

45 基本設計 試作・詳細設計 製造(量産) 設置・調整 運用 エクサ向けアプリケーション開発・利用 システム・ノード設計 実験機 システム製造・設置 P環境設計 P 環境構築 ↓ アプリ開発開始 ↓P 環境更新 8 th Framework Program (FP8) Horizon 2020 第 12 次 5 カ年計画 ↓100 PF 第 13 次 5 カ年計画 1 EF ↓ Development plans of exascale machines in the world (preliminary) 2015/5/3Fifty years of Japanese HPC45

46 9 Key topics for the post-K (Komiyama Committee)  Healthy longevity 1.Innovative drug design 2.Integrated computational bioscience  Disaster mitigation and environment 3.Earthquake and tsunami 4.Advanced prediction of climate and environment 2015/5/3Fifty years of Japanese HPC46

47 9 Key topics for the post-K  Energy problem 5. New technology for energy production, storage and usage 6. Innovative clean energy system  Industry competitiveness 7. New functional devices and materials 8. Innovative design and manufacturing  Basic science 9. Fundamental law and evolution of the universe 2015/5/3Fifty years of Japanese HPC47

48 2015/5/348 Observation of Japanese HPC (1/3)  Japan has manufactured vector supercomputers since 1977  Japanese vector computers were designed as an extension of main frames. –In the early stage, development cost was amortized among mainframes (US-Japan conflicts).  Stress of easiness of use (compilers and tools) –Very good vectorizing compilers. –Users exploited the power of vectorization. –Users were spoiled by them. –Application software development not enough. Fifty years of Japanese HPC

49 Observation of Japanese HPC (2/3)  Until late 1990’s, Japanese vendors focused on vector machines. Late entry to parallel architectures.  Vendors thought parallel machines were for specialized purposes (eg. image processing).  Most users dared not try to harness parallel machines in the 80’s. –Users found difficulties in using message passing.  Some computer scientists were interested in building parallel machines, but they were not used for practical scientific computing. Fifty years of Japanese HPC /5/3

50 Observation of Japanese HPC (3/3)  Practical parallel processing for scientific computing was started by application users: QCDPAX, NWT, CP-PACS, GRAPE’s, ES.  After the K Computer, parallel users increased not only in academia but also in industry  We must prepare for the post-K exascale. –Hierarchical deployment –Home development –Co-design, Science-driven Fifty years of Japanese HPC /5/3

51 Fifty years of Japanese HPC /5/3


Download ppt "2015/5/31 Fifty Years of Japanese HPC ─From transistor to Exascale─ Yoshio Oyanagi Education Center for Computational Sciences Kobe University Fifty years."

Similar presentations


Ads by Google