From FLOPS to BYTES: Disruptive change in high-performance computing towards the post-moore era

Satoshi Matsuoka, Hideharu Amano, Kengo Nakajima, Koji Inoue, Tomohiro Kudoh, Naoya Maruyama, Kenjiro Taura, Takeshi Iwashita, Takahiro Katagiri, Toshihiro Hanawa, Toshio Endo

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Slowdown and inevitable end in exponential scaling of processor performance, the end of the so-called"Moore's Law" is predicted to occur around 2025-2030 timeframe. Because CMOS semiconductor voltage is also approaching its limits, this means that logic transistor power will become constant, and as a result, the system FLOPS will cease to improve, resulting in serious consequences for IT in general, especially supercomputing. Existing attempts to overcome the end of Moore's law are rather limited in their future outlook or applicability. We claim that data-oriented parameters, such as bandwidth and capacity, or BYTES, are the new parameters that will allow continued performance gains for periods even after computing performance or FLOPS ceases to improve, due to continued advances in storage device technologies and optics, and manufacturing technologies including 3-D packaging. Such transition from FLOPS to BYTES will lead to disruptive changes in the overall systems from applications, algorithms, software to architecture, as to what parameter to optimize for, in order to achieve continued performance growth over time. We are launching a new set of research efforts to investigate and devise new technologies to enable such disruptive changes from FLOPS to BYTES in the Post-Moore era, focusing on HPC, where there is extreme sensitivity to performance, and expect the results to disseminate to the rest of IT.

Original languageEnglish
Title of host publication2016 ACM International Conference on Computing Frontiers - Proceedings
PublisherAssociation for Computing Machinery, Inc
Pages274-281
Number of pages8
ISBN (Electronic)9781450341288
DOIs
Publication statusPublished - 2016 May 16
EventACM International Conference on Computing Frontiers, CF 2016 - Como, Italy
Duration: 2016 May 162016 May 18

Other

OtherACM International Conference on Computing Frontiers, CF 2016
CountryItaly
CityComo
Period16/5/1616/5/18

Fingerprint

Launching
Optics
Packaging
Semiconductor materials
Bandwidth
Electric potential
Power transistors

Keywords

  • auto-Tuning
  • deep memory hiearchy. FPGA
  • From BYTES to FLOPS
  • hiearchical algorithms and programming
  • OCS network
  • parallel in space-And-Time
  • Post-Moore
  • post-Moore numerical algorithms
  • postmoore programming model
  • power-constrained supercomputing
  • super building block architecture
  • terabit optics

ASJC Scopus subject areas

  • Software

Cite this

Matsuoka, S., Amano, H., Nakajima, K., Inoue, K., Kudoh, T., Maruyama, N., ... Endo, T. (2016). From FLOPS to BYTES: Disruptive change in high-performance computing towards the post-moore era. In 2016 ACM International Conference on Computing Frontiers - Proceedings (pp. 274-281). Association for Computing Machinery, Inc. https://doi.org/10.1145/2903150.2906830

From FLOPS to BYTES : Disruptive change in high-performance computing towards the post-moore era. / Matsuoka, Satoshi; Amano, Hideharu; Nakajima, Kengo; Inoue, Koji; Kudoh, Tomohiro; Maruyama, Naoya; Taura, Kenjiro; Iwashita, Takeshi; Katagiri, Takahiro; Hanawa, Toshihiro; Endo, Toshio.

2016 ACM International Conference on Computing Frontiers - Proceedings. Association for Computing Machinery, Inc, 2016. p. 274-281.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Matsuoka, S, Amano, H, Nakajima, K, Inoue, K, Kudoh, T, Maruyama, N, Taura, K, Iwashita, T, Katagiri, T, Hanawa, T & Endo, T 2016, From FLOPS to BYTES: Disruptive change in high-performance computing towards the post-moore era. in 2016 ACM International Conference on Computing Frontiers - Proceedings. Association for Computing Machinery, Inc, pp. 274-281, ACM International Conference on Computing Frontiers, CF 2016, Como, Italy, 16/5/16. https://doi.org/10.1145/2903150.2906830
Matsuoka S, Amano H, Nakajima K, Inoue K, Kudoh T, Maruyama N et al. From FLOPS to BYTES: Disruptive change in high-performance computing towards the post-moore era. In 2016 ACM International Conference on Computing Frontiers - Proceedings. Association for Computing Machinery, Inc. 2016. p. 274-281 https://doi.org/10.1145/2903150.2906830
Matsuoka, Satoshi ; Amano, Hideharu ; Nakajima, Kengo ; Inoue, Koji ; Kudoh, Tomohiro ; Maruyama, Naoya ; Taura, Kenjiro ; Iwashita, Takeshi ; Katagiri, Takahiro ; Hanawa, Toshihiro ; Endo, Toshio. / From FLOPS to BYTES : Disruptive change in high-performance computing towards the post-moore era. 2016 ACM International Conference on Computing Frontiers - Proceedings. Association for Computing Machinery, Inc, 2016. pp. 274-281
@inproceedings{d018e588b6754f439f8016b36718fe50,
title = "From FLOPS to BYTES: Disruptive change in high-performance computing towards the post-moore era",
abstract = "Slowdown and inevitable end in exponential scaling of processor performance, the end of the so-called{"}Moore's Law{"} is predicted to occur around 2025-2030 timeframe. Because CMOS semiconductor voltage is also approaching its limits, this means that logic transistor power will become constant, and as a result, the system FLOPS will cease to improve, resulting in serious consequences for IT in general, especially supercomputing. Existing attempts to overcome the end of Moore's law are rather limited in their future outlook or applicability. We claim that data-oriented parameters, such as bandwidth and capacity, or BYTES, are the new parameters that will allow continued performance gains for periods even after computing performance or FLOPS ceases to improve, due to continued advances in storage device technologies and optics, and manufacturing technologies including 3-D packaging. Such transition from FLOPS to BYTES will lead to disruptive changes in the overall systems from applications, algorithms, software to architecture, as to what parameter to optimize for, in order to achieve continued performance growth over time. We are launching a new set of research efforts to investigate and devise new technologies to enable such disruptive changes from FLOPS to BYTES in the Post-Moore era, focusing on HPC, where there is extreme sensitivity to performance, and expect the results to disseminate to the rest of IT.",
keywords = "auto-Tuning, deep memory hiearchy. FPGA, From BYTES to FLOPS, hiearchical algorithms and programming, OCS network, parallel in space-And-Time, Post-Moore, post-Moore numerical algorithms, postmoore programming model, power-constrained supercomputing, super building block architecture, terabit optics",
author = "Satoshi Matsuoka and Hideharu Amano and Kengo Nakajima and Koji Inoue and Tomohiro Kudoh and Naoya Maruyama and Kenjiro Taura and Takeshi Iwashita and Takahiro Katagiri and Toshihiro Hanawa and Toshio Endo",
year = "2016",
month = "5",
day = "16",
doi = "10.1145/2903150.2906830",
language = "English",
pages = "274--281",
booktitle = "2016 ACM International Conference on Computing Frontiers - Proceedings",
publisher = "Association for Computing Machinery, Inc",

}

TY - GEN

T1 - From FLOPS to BYTES

T2 - Disruptive change in high-performance computing towards the post-moore era

AU - Matsuoka, Satoshi

AU - Amano, Hideharu

AU - Nakajima, Kengo

AU - Inoue, Koji

AU - Kudoh, Tomohiro

AU - Maruyama, Naoya

AU - Taura, Kenjiro

AU - Iwashita, Takeshi

AU - Katagiri, Takahiro

AU - Hanawa, Toshihiro

AU - Endo, Toshio

PY - 2016/5/16

Y1 - 2016/5/16

N2 - Slowdown and inevitable end in exponential scaling of processor performance, the end of the so-called"Moore's Law" is predicted to occur around 2025-2030 timeframe. Because CMOS semiconductor voltage is also approaching its limits, this means that logic transistor power will become constant, and as a result, the system FLOPS will cease to improve, resulting in serious consequences for IT in general, especially supercomputing. Existing attempts to overcome the end of Moore's law are rather limited in their future outlook or applicability. We claim that data-oriented parameters, such as bandwidth and capacity, or BYTES, are the new parameters that will allow continued performance gains for periods even after computing performance or FLOPS ceases to improve, due to continued advances in storage device technologies and optics, and manufacturing technologies including 3-D packaging. Such transition from FLOPS to BYTES will lead to disruptive changes in the overall systems from applications, algorithms, software to architecture, as to what parameter to optimize for, in order to achieve continued performance growth over time. We are launching a new set of research efforts to investigate and devise new technologies to enable such disruptive changes from FLOPS to BYTES in the Post-Moore era, focusing on HPC, where there is extreme sensitivity to performance, and expect the results to disseminate to the rest of IT.

AB - Slowdown and inevitable end in exponential scaling of processor performance, the end of the so-called"Moore's Law" is predicted to occur around 2025-2030 timeframe. Because CMOS semiconductor voltage is also approaching its limits, this means that logic transistor power will become constant, and as a result, the system FLOPS will cease to improve, resulting in serious consequences for IT in general, especially supercomputing. Existing attempts to overcome the end of Moore's law are rather limited in their future outlook or applicability. We claim that data-oriented parameters, such as bandwidth and capacity, or BYTES, are the new parameters that will allow continued performance gains for periods even after computing performance or FLOPS ceases to improve, due to continued advances in storage device technologies and optics, and manufacturing technologies including 3-D packaging. Such transition from FLOPS to BYTES will lead to disruptive changes in the overall systems from applications, algorithms, software to architecture, as to what parameter to optimize for, in order to achieve continued performance growth over time. We are launching a new set of research efforts to investigate and devise new technologies to enable such disruptive changes from FLOPS to BYTES in the Post-Moore era, focusing on HPC, where there is extreme sensitivity to performance, and expect the results to disseminate to the rest of IT.

KW - auto-Tuning

KW - deep memory hiearchy. FPGA

KW - From BYTES to FLOPS

KW - hiearchical algorithms and programming

KW - OCS network

KW - parallel in space-And-Time

KW - Post-Moore

KW - post-Moore numerical algorithms

KW - postmoore programming model

KW - power-constrained supercomputing

KW - super building block architecture

KW - terabit optics

UR - http://www.scopus.com/inward/record.url?scp=84978529165&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84978529165&partnerID=8YFLogxK

U2 - 10.1145/2903150.2906830

DO - 10.1145/2903150.2906830

M3 - Conference contribution

AN - SCOPUS:84978529165

SP - 274

EP - 281

BT - 2016 ACM International Conference on Computing Frontiers - Proceedings

PB - Association for Computing Machinery, Inc

ER -