ASIC Clouds: Specializing the Datacenter for Planet-Scale Applications.

Planet-scale applications are driving the exponential growth of the Cloud, and datacenter specialization is the key enabler of this trend. GPU- and FPGA-based clouds have already been deployed to accelerate compute-intensive workloads. ASIC-based clouds are a natural evolution as cloud services expa...

Descripción completa

Detalles Bibliográficos
Publicado en:Communications of the ACM Vol. 63; no. 7; pp. 103 - 110
Autores principales: Taylor, Michael Bedford, Vega, Luis, Khazraee, Moein, Ikuo Magaki, Davidson, Scott, Richmond, Dustin
Formato: Artículo
Publicado: Association for Computing Machinery Jul2020
Materias:
Acceso en línea:Ver este registro en EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=hlh&AN=144780584&site=ehost-live
header:
  @attributes:
    shortDbName: hlh
    uiTerm: 144780584
    longDbName: Humanities International Complete
    uiTag: AN
  controlInfo:
    bkinfo:
    jinfo:
      jid:
        00010782
        ACM
      jtl: Communications of the ACM
      issn: 00010782
      maglogo: N
    pubinfo:
      dt: Jul2020
      vid: 63
      iid: 7
      pid: 68
      pub: Association for Computing Machinery
    artinfo:
      ui:
        144780584
        10.1145/3399734
      ppf: 103
      ppct: 7
      formats:
      tig:
        atl: ASIC Clouds: Specializing the Datacenter for Planet-Scale Applications.
      aug:
        au:
          Taylor, Michael Bedford
          Vega, Luis
          Khazraee, Moein
          Ikuo Magaki
          Davidson, Scott
          Richmond, Dustin
        affil:
          University of Washington, WA, USA
          UC San Diego, CA, USA
      su:
        Application-specific integrated circuits
        Cloud computing
        Central processing units
        Total cost of ownership
        Server farms (Computer network management)
      sug:
        subj:
          Application-specific integrated circuits
          Cloud computing
          Central processing units
          Total cost of ownership
          Server farms (Computer network management)
      ab: Planet-scale applications are driving the exponential growth of the Cloud, and datacenter specialization is the key enabler of this trend. GPU- and FPGA-based clouds have already been deployed to accelerate compute-intensive workloads. ASIC-based clouds are a natural evolution as cloud services expand across the planet. ASIC Clouds are purpose-built datacenters comprised of large arrays of ASIC accelerators that optimize the total cost of ownership (TCO) of large, high-volume scale-out computations. On the surface, ASIC Clouds may seem improbable due to high NREs and ASIC inflexibility, but large-scale ASIC Clouds have already been deployed for the Bitcoin cryptocurrency system. This paper distills lessons from these Bitcoin ASIC Clouds and applies them to other large scale workloads such as YouTube-style video-transcoding and Deep Learning, showing superior TCO versus CPU and GPU. It derives Pareto-optimal ASIC Cloud servers based on accelerator properties, by jointly optimizing ASIC architecture, DRAM, motherboard, power delivery, cooling, and operating voltage. Finally, the authors examine the impact of ASIC NRE and when it makes sense to build an ASIC Cloud.
      pubtype: Periodical
      doctype: Article
      src: R
    language: English
    refInfo:
    copyright:
      @attributes:
        flag: Y
      dt:
        @attributes:
          year: 2020
    holdings:
      @attributes:
        islocal: N