Amano Lab.

Go back to top page

2019

Special Course on Computer Architectures

Slides

  • This class is basically done in 14-203. When the computer room is used, 14-B102 is used.
  • intro_para.pdf /intro.pptx Introduction: Slides used on 4/11
  • All students who take this class is invited in the first day of CoolChips 22 (4/17). http://www.coolchips.org Class in 4/18 is canceled.
  • bsw.pdf /bsw.pptx Bus and Crossbar: Slides used on 4/25
  • snoop.pdf /snoop.pptx Snoop Cache: Slides used on 5/9
  • sync.pdf /sync.pptx Synchronization: Slides used on 5/16
  • numa.pdf /numa.pptx NUMA and directory cache: Slides used on 5/23
  • cmodel.pdf /cmodel.pptx Consistency Models : Slides used on 5/30
  • openmp.pdf /openmp19.tar OpenMP : Slides and files used on 6/6
  • Caution! the lesson on 6/6 will be held in 14-B102. You need Linux account in ITC.
  • mpi.pdf /mpiex19.tar MPI : Slides and files used on 6/13
  • Caution! the lesson on 6/13 will be held in 14-B102.
  • direct.pdf /direct.pptx Direct Interconnection Network : Slides used on 6/20
  • gpu.pdf /cuda19.tar GPU and Cuda: Slides and file used on 6/27
  • reconfhls.pdf testhls.tar VivadoHLS.pdf Reconfigurable Systems: Slides and file used on 7/4
  • Caution! the lesson from 7/11 will be held in 14-203.

  • indirect.pdf /direct.pptx Indirect Interconnection Network : Slides used on 7/11
  • prout.pdf /prout.pptx Packet Routing Methods Slides used on 7/18
  • contest19.pdf /lenet.tar Contest for Cuda Section
  • fpta19.pdf /con19.tar.gz Contest for FPGA Section: Files were updated at 13:00, 7.4. Please load again
  • Caution! Students who did not receive acknowledge mail for the program contest have to inform me asap.

    Parallel Program Contest GPU Section Ranking

  • Sorry. The execution time is not so accurate because of the machine problem.
  • Students who appear here will receive "S". Congratulations!
  • Name Time
    Keisuke Sugiura 0.041
    Sun YuXi 0.042
    Takuya Sakuma 0.044
    Tokio Kihata 0.074
    Ryohei Tomura 0.093
    Masaki Furukawa 0.100
    Keita Suzuki 0.189
    Yoshiaki Honma 0.34
    Kaisei Takahashi 0.76
    Hiroshi Yano 0.9
    Hiroki Hirose 1.1
    Kazuki Masumura 1.1

    Parallel Program Contest FPGA Section Ranking

  • Sorry. The execution time is not so accurate because of the machine problem.
  • Students who appear here will receive "S". Congratulations!
  • Name Time (ms)
    Sun YuXi 0.54
    Hirohisa Watanabe 5.0
    Orsztynowicz Manfred 7.2
    Kohei Itoh 8.5
    Tomoki Shimizu 10.0
    Yuto Kozeki 16.3

    Scoring policy of this class

  • Studunts whose time is smaller than 3ms (GPU) /20ms (FPGA) will receive "A".
  • Other students is ranked with the following standards. Let be the score of parallel programming contest C = 14/T (GPU), 197/T(FPGA). Students can get 0.1 for each exercise, and let be the total score E.
  • C+E > 2.1 : A
  • C+E > 1.2 : B
  • Others: C
  • Note that students whose programs do not work in parallel programming contest cannot receive the unit.

    2018

    Special Course on Computer Architectures

    Parallel Program Contest Ranking

  • I executed the submitted design six times and recorded the best score.
  • Students who appear here will receive "S". Congratulations!
  • Comparc01/02
  • Rank Name Score
    1 Christian Lanius 3.21
    2 Toshiki Kikuchi 2.65
    3 Shun Fujiuchi 2.61
    4 Koki Honda 2.08
    5 Yasuaki Okamoto 1.91
    6 Yosuke Shiiki 1.83

  • Nova
  • Rank Name Score
    1 Takuma Seno 2.64
    2 Shaobai Sun 2.23
    2 Theo Fhaterie 2.23
    4 Kohei Okuoka 2.18
    5 Keisuke Takiguchi 2.04
    6 Yasutake Shigano 1.94
    7' Daisuke Takahashi 1.92
    7 Yuya Ishikawa 1.90
    8 Shoichiro Shohata 1.81

    Scoring policy of this class

  • Studunts whose score is more than 1.44 will receive "A".
  • Other students is ranked with the following standards. Let be the score of parallel programming contest C. Students can get 0.1 for each exercise, and let be the total score E.
  • C+E > 2.1 : A
  • C+E > 1.2 : B
  • Others: C
  • Note that students whose programs do not work in parallel programming contest cannot receive the unit.

    2017

    Parallel Program Contest Ranking

    The followings are temporal results. Students who achieved more than 100 times speedup will receive "S". Congratulations!

    Rank Name Score
    1 Yusuke Takimoto 304.8
    2 Masato Nakada 284.6
    3 Yosuke Fukuchi 259.4
    4 Kit Man Sit 240.2
    5 Katsuhiro Endo 201.4
    6 Motoya Tomoe 195.6
    7 Shu Sekigawa 158.5
    8 Ryosuke Kazami 155.1
    9 Shun Yoshioka 154.5
    10 Tetsushi Toyonaka 154.0
    11 Kensuke Iizuka 126.5
    12 Takuma Iwata 113.4
    13 Mineto Tsukada 110.3
    14 Takumi Hamasaki 106.5
    15 Takeharu Ikezoe 104.7

    2016

    Parallel Program Contest 2016

  • dft.pdf / contest16.tar Slides and toolkit for Contest
  • Parallel Program Contest Ranking

    Thank you for your participation in this contest!
    The score is an average ratio of the execution speed compared to the execution on the CPU.
    Students whose score were more than 10.0 with correct results are recorded here.
    They got "A" regardless of the grades of other execises. Congraturation!

    Rank Name Score
    1 Ryuichi Yamamitsu 328.61
    2 Sho Yuhara 267.26
    3 Atsutoshi Osuka 255.79
    4 Kazusa Musha 173.43
    5 Takuya Kojima 163.05
    6 Koya Mitsuzuka 61.83
    7 Hiroyuki Noda 45.87
    8 Udad Adam 33.29
    9 Kazuma Takemoto 22.08
    10 Kazuki Tanaka 21.14
    11 Tetsui Okubo 19.75
    12 Kenta Ishiguro 18.72
    13 Naoya Takemura 18.72
    14 Laurent Thomas 10.45
    15 Masayuki Takano 10.32

    2015

    Parallel Program Contest Ranking

    Thank you for your participation in this contest!
    The score is an average ratio of the execution speed compared to the execution on the CPU.
    Students whose score were more than 1.5 with correct results are recorded here.
    They got "A" regardless of the grades of other execises. Congraturation!

    Rank Name Score
    1 Daichi Fujiki 30.36
    2 Naoki Ando 23.19
    3 Hideki Shimura 23.16
    4 Jianmei Feng 16.25
    5 Yasuhiro Ono 14.5
    6 Hiroaki Hara 9.78
    7 Ryotaro Sakai 3.01
    8 Kohei Nakamura 2.83
    9 Megumi Okamoto 1.51

    Yasuhiro Sekino's design could only execute 3 of 5 input data sets, but his design was really interesting. Thus, I will give "A" to his design.

    webmaster@am.ics.keio.ac.jp