본문

서브메뉴

Scaling Dataflow: Programmability and Simulation
Scaling Dataflow: Programmability and Simulation
Scaling Dataflow: Programmability and Simulation

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20250211153058
ISBN  
9798346381693
DDC  
005
저자명  
Zhang, Nathan.
서명/저자  
Scaling Dataflow: Programmability and Simulation
발행사항  
[Sl] : Stanford University, 2024
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2024
형태사항  
128 p
주기사항  
Source: Dissertations Abstracts International, Volume: 86-05, Section: A.
주기사항  
Advisor: Olukotun, Oyekunle.
학위논문주기  
Thesis (Ph.D.)--Stanford University, 2024.
초록/해제  
요약In recent years, machine learning has driven massively increased computational demand at all scales - ranging from kiloflop-scale per-packet networking analytics to petaflop-scale large language models (LLMs). Spatial accelerators such as field programmable gate arrays (FPGAs) and other reconfigurable dataflow architectures (RDAs) have emerged as promising candidates, offering significant performance-per-watt improvements over instruction-based architectures such as CPUs and GPUs. However, their widespread adoption is hampered by limitations in programming models and tools. This dissertation tackles two key challenges: (1) constructing efficiently composable libraries for these accelerators, and (2) enabling the parallel simulation of dataflow systems, which are distributed processing units connected by communication channels.Previous work in programming spatial accelerators made writing individual applications significantly easier by lifting the level of abstraction. However, writing single applications is insufficient; reusable libraries are essential to compose larger programs. We first characterize the hierarchical pipeline promotion problem (HPPP), a performance trap which prevents the creation of high performance libraries when implemented with traditional abstractions. We then present streaming tensor interfaces (STIs), a software design pattern which bypasses the HPPP by constructing decoupled streaming pipelines. Using STIs, we construct an efficient library for performing inference, achieving a throughput of 10 cycles-per-inference without batching.We then turn our focus to the simulation of dataflow systems, which sits at the heart of research - from functional testing to performance estimation. Unfortunately, existing parallel software simulation schemes are ill-suited to such systems; optimistic schemes place a large burden on the end-user to reason about undoing speculation errors, while current conservative schemes scale poorly in the face of low-latency high-bandwidth communication. To address these limitations, we introduce the Dataflow Abstract Machine (DAM), a parallel simulator framework which breaks from tradition in both user interface and execution. DAM replaces event-driven modeling with communicating sequential processes, bulk-synchronous time with asynchronous distributed time, and global synchronization with a peer-to-peer protocol. DAM requires 57% less code compared to an existing cycle-based simulation, and outperforms a state-of-the-art simulation framework by at least 2x. Compared to ad-hoc research simulators, DAM achieves speedups of up to five orders of magnitude.Ultimately, both of these advancements are enabled by new abstractions. In this thesis, we will argue that abstractions are key, as abstractions constrain implementations, and implementations dictate performance.
일반주제명  
Programming languages
일반주제명  
Systems design
일반주제명  
Libraries
일반주제명  
Software engineering
일반주제명  
Linear algebra
일반주제명  
Semantics
일반주제명  
Engineers
일반주제명  
Computer engineering
일반주제명  
Computer science
일반주제명  
Design
일반주제명  
Logic
일반주제명  
Systems science
기타저자  
Stanford University.
기본자료저록  
Dissertations Abstracts International. 86-05A.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008250123s2024        us                              c    eng  d
■001000017164878
■00520250211153058
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798346381693
■035    ▼a(MiAaPQ)AAI31652036
■035    ▼a(MiAaPQ)Stanfordbc169pf3686
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a005
■1001  ▼aZhang,  Nathan.
■24510▼aScaling  Dataflow:  Programmability  and  Simulation
■260    ▼a[Sl]▼bStanford  University▼c2024
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2024
■300    ▼a128  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  86-05,  Section:  A.
■500    ▼aAdvisor:  Olukotun,  Oyekunle.
■5021  ▼aThesis  (Ph.D.)--Stanford  University,  2024.
■520    ▼aIn  recent  years,  machine  learning  has  driven  massively  increased  computational  demand  at  all  scales  -  ranging  from  kiloflop-scale  per-packet  networking  analytics  to  petaflop-scale  large  language  models  (LLMs).  Spatial  accelerators  such  as  field  programmable  gate  arrays  (FPGAs)  and  other  reconfigurable  dataflow  architectures  (RDAs)  have  emerged  as  promising  candidates,  offering  significant  performance-per-watt  improvements  over  instruction-based  architectures  such  as  CPUs  and  GPUs.  However,  their  widespread  adoption  is  hampered  by  limitations  in  programming  models  and  tools.  This  dissertation  tackles  two  key  challenges:  (1)  constructing  efficiently  composable  libraries  for  these  accelerators,  and  (2)  enabling  the  parallel  simulation  of  dataflow  systems,  which  are  distributed  processing  units  connected  by  communication  channels.Previous  work  in  programming  spatial  accelerators  made  writing  individual  applications  significantly  easier  by  lifting  the  level  of  abstraction.  However,  writing  single  applications  is  insufficient;  reusable  libraries  are  essential  to  compose  larger  programs.  We  first  characterize  the  hierarchical  pipeline  promotion  problem  (HPPP),  a  performance  trap  which  prevents  the  creation  of  high  performance  libraries  when  implemented  with  traditional  abstractions.  We  then  present  streaming  tensor  interfaces  (STIs),  a  software  design  pattern  which  bypasses  the  HPPP  by  constructing  decoupled  streaming  pipelines.  Using  STIs,  we  construct  an  efficient  library  for  performing  inference,  achieving  a  throughput  of  10  cycles-per-inference  without  batching.We  then  turn  our  focus  to  the  simulation  of  dataflow  systems,  which  sits  at  the  heart  of  research  -  from  functional  testing  to  performance  estimation.  Unfortunately,  existing  parallel  software  simulation  schemes  are  ill-suited  to  such  systems;  optimistic  schemes  place  a  large  burden  on  the  end-user  to  reason  about  undoing  speculation  errors,  while  current  conservative  schemes  scale  poorly  in  the  face  of  low-latency  high-bandwidth  communication.  To  address  these  limitations,  we  introduce  the  Dataflow  Abstract  Machine  (DAM),  a  parallel  simulator  framework  which  breaks  from  tradition  in  both  user  interface  and  execution.  DAM  replaces  event-driven  modeling  with  communicating  sequential  processes,  bulk-synchronous  time  with  asynchronous  distributed  time,  and  global  synchronization  with  a  peer-to-peer  protocol.  DAM  requires  57%  less  code  compared  to  an  existing  cycle-based  simulation,  and  outperforms  a  state-of-the-art  simulation  framework  by  at  least  2x.  Compared  to  ad-hoc  research  simulators,  DAM  achieves  speedups  of  up  to  five  orders  of  magnitude.Ultimately,  both  of  these  advancements  are  enabled  by  new  abstractions.  In  this  thesis,  we  will  argue  that  abstractions  are  key,  as  abstractions  constrain  implementations,  and  implementations  dictate  performance.
■590    ▼aSchool  code:  0212.
■650  4▼aProgramming  languages
■650  4▼aSystems  design
■650  4▼aLibraries
■650  4▼aSoftware  engineering
■650  4▼aLinear  algebra
■650  4▼aSemantics
■650  4▼aEngineers
■650  4▼aComputer  engineering
■650  4▼aComputer  science
■650  4▼aDesign
■650  4▼aLogic
■650  4▼aSystems  science
■690    ▼a0800
■690    ▼a0464
■690    ▼a0984
■690    ▼a0389
■690    ▼a0501
■690    ▼a0395
■690    ▼a0790
■71020▼aStanford  University.
■7730  ▼tDissertations  Abstracts  International▼g86-05A.
■790    ▼a0212
■791    ▼aPh.D.
■792    ▼a2024
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17164878▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

미리보기

내보내기

chatGPT토론

Ai 추천 관련 도서


    신착도서 더보기
    최근 3년간 통계입니다.

    소장정보

    • 예약
    • 소재불명신고
    • 나의폴더
    • 우선정리요청
    • 비도서대출신청
    • 야간 도서대출신청
    소장자료
    등록번호 청구기호 소장처 대출가능여부 대출정보
    TF12476 전자도서 대출가능 마이폴더 부재도서신고 비도서대출신청 야간 도서대출신청

    * 대출중인 자료에 한하여 예약이 가능합니다. 예약을 원하시면 예약버튼을 클릭하십시오.

    해당 도서를 다른 이용자가 함께 대출한 도서

    관련 인기도서

    로그인 후 이용 가능합니다.