PrefixRL: Optimization of Parallel Prefix Circuits using Deep Reinforcement Learning

In this work, we present a reinforcement learning (RL) based approach to designing parallel prefix circuits such as adders or priority encoders that are fundamental to high-performance digital design. Unlike prior methods, our approach designs solutions tabula rasa purely through learning with synth...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:2021 58th ACM/IEEE Design Automation Conference (DAC) s. 853 - 858
Hlavní autori: Roy, Rajarshi, Raiman, Jonathan, Kant, Neel, Elkin, Ilyas, Kirby, Robert, Siu, Michael, Oberman, Stuart, Godil, Saad, Catanzaro, Bryan
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 05.12.2021
Predmet:
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract In this work, we present a reinforcement learning (RL) based approach to designing parallel prefix circuits such as adders or priority encoders that are fundamental to high-performance digital design. Unlike prior methods, our approach designs solutions tabula rasa purely through learning with synthesis in the loop. We design a grid-based state-action representation and an RL environment for constructing legal prefix circuits. Deep Convolutional RL agents trained on this environment produce prefix adder circuits that Pareto-dominate existing baselines with up to 16.0% and 30.2% lower area for the same delay in the 32b and 64b settings respectively. We observe that agents trained with open-source synthesis tools and cell library can design adder circuits that achieve lower area and delay than commercial tool adders in an industrial cell library.
AbstractList In this work, we present a reinforcement learning (RL) based approach to designing parallel prefix circuits such as adders or priority encoders that are fundamental to high-performance digital design. Unlike prior methods, our approach designs solutions tabula rasa purely through learning with synthesis in the loop. We design a grid-based state-action representation and an RL environment for constructing legal prefix circuits. Deep Convolutional RL agents trained on this environment produce prefix adder circuits that Pareto-dominate existing baselines with up to 16.0% and 30.2% lower area for the same delay in the 32b and 64b settings respectively. We observe that agents trained with open-source synthesis tools and cell library can design adder circuits that achieve lower area and delay than commercial tool adders in an industrial cell library.
Author Raiman, Jonathan
Oberman, Stuart
Roy, Rajarshi
Catanzaro, Bryan
Godil, Saad
Siu, Michael
Kant, Neel
Elkin, Ilyas
Kirby, Robert
Author_xml – sequence: 1
  givenname: Rajarshi
  surname: Roy
  fullname: Roy, Rajarshi
  email: rajarshir@nvidia.com
  organization: NVIDIA,Santa Clara,CA,USA
– sequence: 2
  givenname: Jonathan
  surname: Raiman
  fullname: Raiman, Jonathan
  email: jraiman@nvidia.com
  organization: NVIDIA,Santa Clara,CA,USA
– sequence: 3
  givenname: Neel
  surname: Kant
  fullname: Kant, Neel
  email: nkant@nvidia.com
  organization: NVIDIA,Santa Clara,CA,USA
– sequence: 4
  givenname: Ilyas
  surname: Elkin
  fullname: Elkin, Ilyas
  email: ielkin@nvidia.com
  organization: NVIDIA,Santa Clara,CA,USA
– sequence: 5
  givenname: Robert
  surname: Kirby
  fullname: Kirby, Robert
  email: rkirby@nvidia.com
  organization: NVIDIA,Santa Clara,CA,USA
– sequence: 6
  givenname: Michael
  surname: Siu
  fullname: Siu, Michael
  email: msiu@nvidia.com
  organization: NVIDIA,Santa Clara,CA,USA
– sequence: 7
  givenname: Stuart
  surname: Oberman
  fullname: Oberman, Stuart
  email: soberman@nvidia.com
  organization: NVIDIA,Santa Clara,CA,USA
– sequence: 8
  givenname: Saad
  surname: Godil
  fullname: Godil, Saad
  email: sgodil@nvidia.com
  organization: NVIDIA,Santa Clara,CA,USA
– sequence: 9
  givenname: Bryan
  surname: Catanzaro
  fullname: Catanzaro, Bryan
  email: bcatanzaro@nvidia.com
  organization: NVIDIA,Santa Clara,CA,USA
BookMark eNotj8tKw0AYhUdQUGueQIR5gca5X9yV1BsEWkpdl5nkHxlIJmGSgvr0FtrNOYvv8MG5R9dpSIDQEyUlpcQ-r1cVNUSLkhFGSyuNIlZcocJqQ5WSgjMtyC0qpil6oog04pR3aL_NEOLPrn7Bm3GOffxzcxwSHgLeuuy6Djp8nuAq5uYY5wkfp5i-8RpgxDuIKQy5gR7SjGtwOZ3YA7oJrpuguPQCfb297quPZb15_6xW9dIxo-elc5o76w2AbKz3IbS8FZxyCmC0U1waZmwrvfLSBKASmGWiNWCt19QSxhfo8eyNAHAYc-xd_j1cvvN_ePRShA
ContentType Conference Proceeding
DBID 6IE
6IH
CBEJK
RIE
RIO
DOI 10.1109/DAC18074.2021.9586094
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan (POP) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP) 1998-present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
EISBN 9781665432740
1665432748
EndPage 858
ExternalDocumentID 9586094
Genre orig-research
GroupedDBID 6IE
6IH
ACM
ALMA_UNASSIGNED_HOLDINGS
CBEJK
RIE
RIO
ID FETCH-LOGICAL-a287t-aa73a9b8ee5c9bbffd3d43131ee87a6358289d5b6b58fe15e2924d8e99b719023
IEDL.DBID RIE
ISICitedReferencesCount 32
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000766079700143&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
IngestDate Wed Aug 27 02:28:30 EDT 2025
IsPeerReviewed false
IsScholarly true
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-a287t-aa73a9b8ee5c9bbffd3d43131ee87a6358289d5b6b58fe15e2924d8e99b719023
PageCount 6
ParticipantIDs ieee_primary_9586094
PublicationCentury 2000
PublicationDate 2021-Dec.-5
PublicationDateYYYYMMDD 2021-12-05
PublicationDate_xml – month: 12
  year: 2021
  text: 2021-Dec.-5
  day: 05
PublicationDecade 2020
PublicationTitle 2021 58th ACM/IEEE Design Automation Conference (DAC)
PublicationTitleAbbrev DAC
PublicationYear 2021
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssib060584060
Score 2.4378808
Snippet In this work, we present a reinforcement learning (RL) based approach to designing parallel prefix circuits such as adders or priority encoders that are...
SourceID ieee
SourceType Publisher
StartPage 853
SubjectTerms datapath optimization
Delays
Libraries
machine learning
Machine learning algorithms
Reinforcement learning
Task analysis
Training
Title PrefixRL: Optimization of Parallel Prefix Circuits using Deep Reinforcement Learning
URI https://ieeexplore.ieee.org/document/9586094
WOSCitedRecordID wos000766079700143&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1NSwMxEA1t8eBJpRW_ycGj226aZpN4k9biQepSKvRWkuykLGhbtrvizzdJ14rgxVsI-YBJwuQl894gdKssZ9oMdCSzxAEUyuNIE6OjRDs8pHWsNLUh2QSfTMR8LtMGuttzYQAgBJ9B1xfDX362NpV_KutJJhIHR5qoyXmy42p97x3_u-d8U1yTdEgse6OHIfFSLw4E9km37vsriUrwIeOj_81-jDo_ZDyc7t3MCWrAqo1mqfNs-ef0-R6_uEP_XrMp8driVBU-P8ob3jXBw7wwVV5usQ9xX-IRwAZPIQimmvA2iGuN1WUHvY4fZ8OnqE6QECkHdMpIKU6V1AKAGam1tRnN3IWAEgDBVeJJsEJmTCeaCQuEQd-hrUyAlJq7i0CfnqLWar2CM4QTbhW1SsaMmcHAjZaRGCin0uEPr2B4jtreIovNTgNjURvj4u_qS3TojR7CPtgVapVFBdfowHyU-ba4CQv3BYgumos
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3PS8MwGA1zCnpS2cTf5uDRbu2yNIk32ZSJc5YxYbeRpF9GQbfRdeKfb5LVieDFWwhpAknDl5d87z2ErqVhVOm2CkQaW4BCWBioSKsgVhYPKRVKRYw3m2CDAR-PRVJBNxsuDAD45DNouKJ_y0_neuWuypqC8tjCkS207ZyzSrbW99_j3vdsdApLmk4Uimb3rhM5sRcLA1tRo_z6l42KjyIP-_8b_wDVf-h4ONkEmkNUgVkNjRIb27LPYf8Wv9ht_17yKfHc4ETmziHlDa-b4E6W61VWLLFLcp_iLsACD8FLpmp_O4hLldVpHb0-3I86vaC0SAikhTpFICUjUigOQLVQypiUpPZIQCIAzmTsaLBcpFTFinIDEYWWxVspByEUs0eBFjlC1dl8BscIx8xIYqQIKdXttu0tjUIgjAiLQJyG4QmquRmZLNYqGJNyMk7_rr5Cu73Rc3_Sfxw8naE9twA-CYSeo2qRr-AC7eiPIlvml34RvwDGQp3U
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=2021+58th+ACM%2FIEEE+Design+Automation+Conference+%28DAC%29&rft.atitle=PrefixRL%3A+Optimization+of+Parallel+Prefix+Circuits+using+Deep+Reinforcement+Learning&rft.au=Roy%2C+Rajarshi&rft.au=Raiman%2C+Jonathan&rft.au=Kant%2C+Neel&rft.au=Elkin%2C+Ilyas&rft.date=2021-12-05&rft.pub=IEEE&rft.spage=853&rft.epage=858&rft_id=info:doi/10.1109%2FDAC18074.2021.9586094&rft.externalDocID=9586094