Wafer-level multi-GPU simulation method and device based on Gem5 framework and storage medium

The invention discloses a multi-GPU (Graphics Processing Unit) analogue simulation method based on a Gem5 framework, which is used for a wafer-level system and comprises the following steps of: building a Gem5 system simulation running environment, downloading a source file, compiling and running; t...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: LIU ZHANWEI, PEI XUE, YU HONG, SHAO YANGXUE, LI PEIJIE, WEI SHUAI, WANG SINING
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator LIU ZHANWEI
PEI XUE
YU HONG
SHAO YANGXUE
LI PEIJIE
WEI SHUAI
WANG SINING
description The invention discloses a multi-GPU (Graphics Processing Unit) analogue simulation method based on a Gem5 framework, which is used for a wafer-level system and comprises the following steps of: building a Gem5 system simulation running environment, downloading a source file, compiling and running; the method comprises the following steps: creating a plurality of GPU components, respectively connecting L2 caches of a plurality of GPUs to a Directory, and respectively pointing the L2 caches to memory segments to which the GPUs belong by the Directory; the ROCk driver is changed, a plurality of software queues are created for one or more programs, the software queues are distributed to a plurality of GPUs, and a doorbell region is created for each GPU; and carrying out simulation test on the wafer-level system of the multiple GPUs by adopting the built Gem5 system, and verifying the system performance. Therefore, communication, interaction and scheduling among multiple GPUs are realized, so that the design time
format Patent
fullrecord <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_CN116776821A</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>CN116776821A</sourcerecordid><originalsourceid>FETCH-epo_espacenet_CN116776821A3</originalsourceid><addsrcrecordid>eNqNi7EOgjAURbs4GPUfnh_QAY3AaojiZBw0ToY86UUbW0ragr8vMX6A0z3JOXcqbldu4KXBAEO2N1HL8nShoEfmqF1LFvHpFHGrSGHQNejOAYpGVcJuqPFs8Xb-9U1CdJ4fGF9K93YuJg2bgMVvZ2K5352Lg0TnKoSOa7SIVXFMkjTL0nyVbNf_NB98_Tta</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Wafer-level multi-GPU simulation method and device based on Gem5 framework and storage medium</title><source>esp@cenet</source><creator>LIU ZHANWEI ; PEI XUE ; YU HONG ; SHAO YANGXUE ; LI PEIJIE ; WEI SHUAI ; WANG SINING</creator><creatorcontrib>LIU ZHANWEI ; PEI XUE ; YU HONG ; SHAO YANGXUE ; LI PEIJIE ; WEI SHUAI ; WANG SINING</creatorcontrib><description>The invention discloses a multi-GPU (Graphics Processing Unit) analogue simulation method based on a Gem5 framework, which is used for a wafer-level system and comprises the following steps of: building a Gem5 system simulation running environment, downloading a source file, compiling and running; the method comprises the following steps: creating a plurality of GPU components, respectively connecting L2 caches of a plurality of GPUs to a Directory, and respectively pointing the L2 caches to memory segments to which the GPUs belong by the Directory; the ROCk driver is changed, a plurality of software queues are created for one or more programs, the software queues are distributed to a plurality of GPUs, and a doorbell region is created for each GPU; and carrying out simulation test on the wafer-level system of the multiple GPUs by adopting the built Gem5 system, and verifying the system performance. Therefore, communication, interaction and scheduling among multiple GPUs are realized, so that the design time</description><language>chi ; eng</language><subject>CALCULATING ; COMPUTING ; COUNTING ; ELECTRIC DIGITAL DATA PROCESSING ; PHYSICS</subject><creationdate>2023</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20230919&amp;DB=EPODOC&amp;CC=CN&amp;NR=116776821A$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25555,76308</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20230919&amp;DB=EPODOC&amp;CC=CN&amp;NR=116776821A$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>LIU ZHANWEI</creatorcontrib><creatorcontrib>PEI XUE</creatorcontrib><creatorcontrib>YU HONG</creatorcontrib><creatorcontrib>SHAO YANGXUE</creatorcontrib><creatorcontrib>LI PEIJIE</creatorcontrib><creatorcontrib>WEI SHUAI</creatorcontrib><creatorcontrib>WANG SINING</creatorcontrib><title>Wafer-level multi-GPU simulation method and device based on Gem5 framework and storage medium</title><description>The invention discloses a multi-GPU (Graphics Processing Unit) analogue simulation method based on a Gem5 framework, which is used for a wafer-level system and comprises the following steps of: building a Gem5 system simulation running environment, downloading a source file, compiling and running; the method comprises the following steps: creating a plurality of GPU components, respectively connecting L2 caches of a plurality of GPUs to a Directory, and respectively pointing the L2 caches to memory segments to which the GPUs belong by the Directory; the ROCk driver is changed, a plurality of software queues are created for one or more programs, the software queues are distributed to a plurality of GPUs, and a doorbell region is created for each GPU; and carrying out simulation test on the wafer-level system of the multiple GPUs by adopting the built Gem5 system, and verifying the system performance. Therefore, communication, interaction and scheduling among multiple GPUs are realized, so that the design time</description><subject>CALCULATING</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>ELECTRIC DIGITAL DATA PROCESSING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2023</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNqNi7EOgjAURbs4GPUfnh_QAY3AaojiZBw0ToY86UUbW0ragr8vMX6A0z3JOXcqbldu4KXBAEO2N1HL8nShoEfmqF1LFvHpFHGrSGHQNejOAYpGVcJuqPFs8Xb-9U1CdJ4fGF9K93YuJg2bgMVvZ2K5352Lg0TnKoSOa7SIVXFMkjTL0nyVbNf_NB98_Tta</recordid><startdate>20230919</startdate><enddate>20230919</enddate><creator>LIU ZHANWEI</creator><creator>PEI XUE</creator><creator>YU HONG</creator><creator>SHAO YANGXUE</creator><creator>LI PEIJIE</creator><creator>WEI SHUAI</creator><creator>WANG SINING</creator><scope>EVB</scope></search><sort><creationdate>20230919</creationdate><title>Wafer-level multi-GPU simulation method and device based on Gem5 framework and storage medium</title><author>LIU ZHANWEI ; PEI XUE ; YU HONG ; SHAO YANGXUE ; LI PEIJIE ; WEI SHUAI ; WANG SINING</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_CN116776821A3</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>chi ; eng</language><creationdate>2023</creationdate><topic>CALCULATING</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>ELECTRIC DIGITAL DATA PROCESSING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>LIU ZHANWEI</creatorcontrib><creatorcontrib>PEI XUE</creatorcontrib><creatorcontrib>YU HONG</creatorcontrib><creatorcontrib>SHAO YANGXUE</creatorcontrib><creatorcontrib>LI PEIJIE</creatorcontrib><creatorcontrib>WEI SHUAI</creatorcontrib><creatorcontrib>WANG SINING</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>LIU ZHANWEI</au><au>PEI XUE</au><au>YU HONG</au><au>SHAO YANGXUE</au><au>LI PEIJIE</au><au>WEI SHUAI</au><au>WANG SINING</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Wafer-level multi-GPU simulation method and device based on Gem5 framework and storage medium</title><date>2023-09-19</date><risdate>2023</risdate><abstract>The invention discloses a multi-GPU (Graphics Processing Unit) analogue simulation method based on a Gem5 framework, which is used for a wafer-level system and comprises the following steps of: building a Gem5 system simulation running environment, downloading a source file, compiling and running; the method comprises the following steps: creating a plurality of GPU components, respectively connecting L2 caches of a plurality of GPUs to a Directory, and respectively pointing the L2 caches to memory segments to which the GPUs belong by the Directory; the ROCk driver is changed, a plurality of software queues are created for one or more programs, the software queues are distributed to a plurality of GPUs, and a doorbell region is created for each GPU; and carrying out simulation test on the wafer-level system of the multiple GPUs by adopting the built Gem5 system, and verifying the system performance. Therefore, communication, interaction and scheduling among multiple GPUs are realized, so that the design time</abstract><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier
ispartof
issn
language chi ; eng
recordid cdi_epo_espacenet_CN116776821A
source esp@cenet
subjects CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
title Wafer-level multi-GPU simulation method and device based on Gem5 framework and storage medium
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-15T03%3A40%3A14IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=LIU%20ZHANWEI&rft.date=2023-09-19&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3ECN116776821A%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true