Analysis of Model Parallelism for AI Applications on a 64-core RV64 Server CPU

Massive Data Parallel workloads, driven by inference on large ML models, are pushing hardware vendors to develop efficient and cost-effective multi-core server CPUs. The RISC-V architecture plays a prominent role due to its open, extensible, and energy-friendly ISA. Despite significant progress in r...

Celý popis

Uložené v:

Podrobná bibliografia
Vydané v:	International journal of parallel programming Ročník 53; číslo 4; s. 27
Hlavní autori:	Malenza, Giulio, Garcia, Adriano Marques, Birke, Robert, Benini, Luca, Aldinucci, Marco
Médium:	Journal Article
Jazyk:	English
Vydavateľské údaje:	New York Springer US 01.08.2025 Springer Nature B.V
Predmet:	Artificial intelligence Central processing units Computer Science Computer vision CPUs Design Fashion models Hardware Inference Machine learning Neural networks Open standards Parallel processing Pipelining (computers) Processor Architectures Servers Software Software Engineering/Programming and Operating Systems Theory of Computation Workloads AI RISC-V Model parallelism PyTorch SOPHON SG2042
ISSN:	0885-7458, 1573-7640
On-line prístup:	Získať plný text
Tagy:	Pridať tag Žiadne tagy, Buďte prvý, kto otaguje tento záznam!

Abstract	Massive Data Parallel workloads, driven by inference on large ML models, are pushing hardware vendors to develop efficient and cost-effective multi-core server CPUs. The RISC-V architecture plays a prominent role due to its open, extensible, and energy-friendly ISA. Despite significant progress in recent years, finding efficient methods to run AI applications in parallel on new architectures to fully harness their maximum performance remains a challenge. In this study, we investigate the impact of model parallelism on the inference of machine learning models on the SOPHON SG2042 SoC, the first server-grade CPU based on the RV64 ISA, composed of 64 cores arranged in a grid of 16 groups of 4 cores. Specifically, we aim to enhance performance via better data locality stemming from splitting and assigning parts of the model to specific (groups of) cores handling dependencies via a pipeline execution. We orchestrate execution using FastFlow, a low-level programming framework designed for multithreaded streaming applications. By comparing the results against the standard multi-core inference approach based on data parallelism and analyzing the effects of different submodel-to-core mapping strategies, we aim to provide a comprehensive understanding of how the model parallel approach can maximize efficiency and utilization of hardware resources. In our experiments, using model parallelism improved up to 8.4 times the performance over the native PyTorch parallelism.
AbstractList	Massive Data Parallel workloads, driven by inference on large ML models, are pushing hardware vendors to develop efficient and cost-effective multi-core server CPUs. The RISC-V architecture plays a prominent role due to its open, extensible, and energy-friendly ISA. Despite significant progress in recent years, finding efficient methods to run AI applications in parallel on new architectures to fully harness their maximum performance remains a challenge. In this study, we investigate the impact of model parallelism on the inference of machine learning models on the SOPHON SG2042 SoC, the first server-grade CPU based on the RV64 ISA, composed of 64 cores arranged in a grid of 16 groups of 4 cores. Specifically, we aim to enhance performance via better data locality stemming from splitting and assigning parts of the model to specific (groups of) cores handling dependencies via a pipeline execution. We orchestrate execution using FastFlow, a low-level programming framework designed for multithreaded streaming applications. By comparing the results against the standard multi-core inference approach based on data parallelism and analyzing the effects of different submodel-to-core mapping strategies, we aim to provide a comprehensive understanding of how the model parallel approach can maximize efficiency and utilization of hardware resources. In our experiments, using model parallelism improved up to 8.4 times the performance over the native PyTorch parallelism.
ArticleNumber	27
Author	Malenza, Giulio Aldinucci, Marco Birke, Robert Garcia, Adriano Marques Benini, Luca
Author_xml	– sequence: 1 givenname: Giulio surname: Malenza fullname: Malenza, Giulio email: giulio.malenza@unito.it organization: Department of Computer Science, University of Turin – sequence: 2 givenname: Adriano Marques surname: Garcia fullname: Garcia, Adriano Marques email: adriano.marquesgarcia@unito.it organization: Department of Computer Science, University of Turin – sequence: 3 givenname: Robert surname: Birke fullname: Birke, Robert email: robert.birke@unito.it organization: Department of Computer Science, University of Turin – sequence: 4 givenname: Luca surname: Benini fullname: Benini, Luca organization: Department of Computer Science, University of Bologna – sequence: 5 givenname: Marco surname: Aldinucci fullname: Aldinucci, Marco organization: Department of Computer Science, University of Turin
BookMark	eNp9kEtLAzEURoNUsK3-AVcB19GbZPKYZSk-ClWLWrchM8nIlOlkTFqh_96pI7hzdTfnfHDPBI3a0HqELilcUwB1kygoKQkwQQA0MCJP0JgKxYmSGYzQGLQWRGVCn6FJShsAyJXWY_Q0a21zSHXCocKPwfkGr2y0TeObOm1xFSKeLfCs65q6tLs6tD3YYotlRsoQPX55lxl-9fHLRzxfrc_RaWWb5C9-7xSt727f5g9k-Xy_mM-WpGSK7YizvBJ5roXW1BVW55mrcq-0gMyBKlhVOOpAcyi1tIJX_W_Kel70EuOly_kUXQ27XQyfe592ZhP2sX8lGc6YkJwpeqTYQJUxpBR9ZbpYb208GArm2M0M3Uzfzfx0M7KX-CClHm4_fPyb_sf6Biufb4c
Cites_doi	10.1007/978-1-4842-4398-5 10.1007/s11227-022-05024-y 10.21236/ADA605735 10.1109/CVPR.2016.90 10.1002/9781119332015.ch13 10.1145/3065386 10.1109/MM.2024.3373763 10.1145/2503210.2503219 10.1002/cpe.4175 10.1007/s11227-024-05934-z 10.1145/3587135.3592211 10.1109/CVPR.2009.5206848 10.1145/3203217.3205340 10.1145/3624062.3624234 10.1007/978-3-031-40843-4_31 10.1145/3528416.3530869 10.1109/72.279181 10.1016/0167-739X(92)90040-I 10.3390/info14020064 10.1109/ISCA45697.2020.00016 10.1007/s10766-022-00750-5
ContentType	Journal Article
Copyright	The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025 Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.
Copyright_xml	– notice: The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025 Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. – notice: The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.
DBID	AAYXX CITATION 7SC 8FD JQ2 L7M L~C L~D
DOI	10.1007/s10766-025-00802-6
DatabaseName	CrossRef Computer and Information Systems Abstracts Technology Research Database ProQuest Computer Science Collection Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Academic Computer and Information Systems Abstracts Professional
DatabaseTitle	CrossRef Computer and Information Systems Abstracts Technology Research Database Computer and Information Systems Abstracts – Academic Advanced Technologies Database with Aerospace ProQuest Computer Science Collection Computer and Information Systems Abstracts Professional
DatabaseTitleList	Computer and Information Systems Abstracts
DeliveryMethod	fulltext_linktorsrc
Discipline	Computer Science
EISSN	1573-7640
ExternalDocumentID	10_1007_s10766_025_00802_6
GrantInformation_xml	– fundername: DYMAN grantid: 101161930 – fundername: “FutureHPC & BigData” of the ICSC - Centro Nazionale di Ricerca in “High Performance Computing, Big Data and Quantum Computing
GroupedDBID	-~X .86 .DC .VR 06D 0R~ 0VY 199 1N0 203 29J 2J2 2JN 2JY 2KG 2LR 2~H 30V 4.4 406 408 409 40D 40E 5GY 5VS 67Z 6NX 78A 7WY 8FL 8TC 8UJ 95- 95. 95~ 96X AABHQ AACDK AAHNG AAIAL AAJBT AAJKR AANZL AAPKM AARTL AASML AATNV AATVU AAUYE AAWCG AAYIU AAYQN AAYZH ABAKF ABBBX ABBRH ABBXA ABDBE ABDBF ABDZT ABECU ABFSG ABFTD ABFTV ABHLI ABHQN ABJNI ABJOX ABKCH ABKTR ABMNI ABMQK ABNWP ABQBU ABRTQ ABSXP ABTEG ABTHY ABTKH ABTMW ABWNU ABXPI ACAOD ACDTI ACGFO ACGFS ACHSB ACHXU ACIHN ACKNC ACMDZ ACMLO ACNCT ACOKC ACOMO ACPIV ACREN ACSTC ACUHS ACZOJ ADHIR ADKNI ADKPE ADRFC ADTPH ADURQ ADYFF ADYOE ADZKW AEAQA AEFQL AEGAL AEGNC AEJHL AEJRE AEMSY AENEX AEOHA AEPYU AESKC AETLH AEVLU AEXYK AEZWR AFBBN AFDZB AFHIU AFLOW AFOHR AFQWF AFWTZ AFYQB AFZKB AGAYW AGDGC AGJBK AGMZJ AGQEE AGQMX AGRTI AGWIL AGWZB AGYKE AHAVH AHBYD AHKAY AHPBZ AHSBF AHWEU AHYZX AIAKS AIGIU AIIXL AILAN AITGF AIXLP AJRNO AJZVZ ALMA_UNASSIGNED_HOLDINGS ALWAN AMKLP AMTXH AMXSW AMYLF AOCGG ARAPS ARMRJ ATHPR AXYYD AYFIA AYJHY AZFZN B-. BA0 BENPR BGNMA BKOMP BSONS CS3 CSCUP DDRTE DL5 DNIVK DPUIP DU5 EAP EBLON EBS EIOEI ESBYG ESX FEDTE FERAY FFXSO FIGPU FNLPD FRRFC FWDCC GGCAI GGRSB GJIRD GNWQR GQ7 GQ8 GXS HCIFZ HF~ HG5 HG6 HMJXF HQYDN HRMNR HVGLF HZ~ I09 IHE IJ- IKXTQ ITM IWAJR IXC IZIGR IZQ I~X I~Z J-C J0Z JBSCW JCJTX JZLTJ K60 K6~ KDC KOV LAK LLZTM M2O M4Y MA- MS~ NB0 NPVJJ NQJWS NU0 O93 O9G O9I O9J OAM P19 P9O PF0 PT4 PT5 QOK QOS R89 R9I RHV RNS ROL RPX RSV S16 S1Z S27 S3B SAP SCO SDH SDM SHX SISQX SJYHP SNE SNPRN SNX SOHCF SOJ SPISZ SRMVM SSLCW STPWE SZN T13 TAE TN5 TSG TSK TSV TUC U2A U5U UG4 UOJIU UTJUX VC2 W23 W48 WH7 WK8 YLTOR Z45 ZMTXR ~EX -Y2 .4S 2.D 28- 2P1 2VQ 5QI 8FE 8FG 8G5 AAOBN AARHV AAYJJ AAYTO AAYXX ABDPE ABFSI ABQSL ABULA ABUWG ACBXY ADHKG ADMLS AEBTG AEFIE AEKMD AFEXP AFFHD AFGCZ AFKRA AGGDS AGQPQ AJBLW ARCSS AZQEC B0M BBWZM BDATZ BEZIV BGLVJ BPHCQ CAG CCPQU CITATION COF DWQXO E.L EAD EAS EDO EJD EMK EPL FINBP FRNLG FSGXE GNUQQ GROUPED_ABI_INFORM_RESEARCH GUQSH H13 H~9 I-F K6V K7- KOW M0C N2Q NDZJH O9- OVD P62 PHGZM PHGZT PQBIZ PQBZA PQGLB PQQKQ PROAC Q2X RNI RZC RZE RZK S26 S28 SCJ SCLPG T16 TEORI TUS UZXMN VFIZW ZY4 ~8M 7SC 8FD JQ2 L7M L~C L~D
ID	FETCH-LOGICAL-c272t-da3f59985881dba894df9e78504d07b2fbd1d0830c86a53f7667ae3bf5923cd93
IEDL.DBID	RSV
ISICitedReferencesCount	0
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001520151400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	0885-7458
IngestDate	Wed Nov 05 09:29:50 EST 2025 Sat Nov 29 07:35:22 EST 2025 Sat Jul 26 01:16:00 EDT 2025
IsPeerReviewed	true
IsScholarly	true
Issue	4
Keywords	AI RISC-V Model parallelism PyTorch SOPHON SG2042
Language	English
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c272t-da3f59985881dba894df9e78504d07b2fbd1d0830c86a53f7667ae3bf5923cd93
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14
PQID	3225632719
PQPubID	48389
ParticipantIDs	proquest_journals_3225632719 crossref_primary_10_1007_s10766_025_00802_6 springer_journals_10_1007_s10766_025_00802_6
PublicationCentury	2000
PublicationDate	2025-08-01
PublicationDateYYYYMMDD	2025-08-01
PublicationDate_xml	– month: 08 year: 2025 text: 2025-08-01 day: 01
PublicationDecade	2020
PublicationPlace	New York
PublicationPlace_xml	– name: New York
PublicationTitle	International journal of parallel programming
PublicationTitleAbbrev	Int J Parallel Prog
PublicationYear	2025
Publisher	Springer US Springer Nature B.V
Publisher_xml	– name: Springer US – name: Springer Nature B.V
References	802_CR27 802_CR29 802_CR28 802_CR21 802_CR20 802_CR23 JKL Lee (802_CR37) 2023 802_CR22 802_CR25 802_CR6 802_CR7 802_CR4 FG Zee (802_CR32) 2015; 41 802_CR5 802_CR8 D del Rio Astorga (802_CR15) 2017; 29 AM Garcia (802_CR17) 2023; 79 802_CR16 802_CR38 802_CR18 802_CR39 N Tonci (802_CR9) 2022; 51 802_CR19 A Krizhevsky (802_CR24) 2017; 60 802_CR30 802_CR10 802_CR31 802_CR2 802_CR12 802_CR34 802_CR3 802_CR11 802_CR33 802_CR14 802_CR36 S Kalapothas (802_CR1) 2023 802_CR35 M Danelutto (802_CR13) 1992; 8 Y Bengio (802_CR26) 1994; 5
References_xml	– ident: 802_CR16 doi: 10.1007/978-1-4842-4398-5 – volume: 79 start-page: 9206 issue: 8 year: 2023 ident: 802_CR17 publication-title: J. Supercomput. doi: 10.1007/s11227-022-05024-y – ident: 802_CR30 – ident: 802_CR6 doi: 10.21236/ADA605735 – ident: 802_CR11 – ident: 802_CR7 doi: 10.1109/CVPR.2016.90 – ident: 802_CR8 doi: 10.1002/9781119332015.ch13 – volume: 60 start-page: 84 issue: 6 year: 2017 ident: 802_CR24 publication-title: Commun. ACM doi: 10.1145/3065386 – ident: 802_CR4 doi: 10.1109/MM.2024.3373763 – ident: 802_CR20 – ident: 802_CR3 – ident: 802_CR22 – ident: 802_CR19 doi: 10.1145/2503210.2503219 – volume: 29 start-page: 4175 issue: 24 year: 2017 ident: 802_CR15 publication-title: Concurr. Comput. Pract. Exp. doi: 10.1002/cpe.4175 – ident: 802_CR14 doi: 10.1007/s11227-024-05934-z – ident: 802_CR34 doi: 10.1145/3587135.3592211 – ident: 802_CR33 – ident: 802_CR10 – ident: 802_CR31 – ident: 802_CR12 – ident: 802_CR28 doi: 10.1109/CVPR.2009.5206848 – ident: 802_CR18 – ident: 802_CR39 – ident: 802_CR21 – ident: 802_CR36 doi: 10.1145/3203217.3205340 – ident: 802_CR2 – ident: 802_CR23 – volume: 41 start-page: 14 issue: 3 year: 2015 ident: 802_CR32 publication-title: ACM Trans. Math. Softw. – ident: 802_CR38 doi: 10.1145/3624062.3624234 – start-page: 419 volume-title: High Performance Computing year: 2023 ident: 802_CR37 doi: 10.1007/978-3-031-40843-4_31 – ident: 802_CR27 – ident: 802_CR35 doi: 10.1145/3528416.3530869 – ident: 802_CR25 – ident: 802_CR29 – volume: 5 start-page: 157 issue: 2 year: 1994 ident: 802_CR26 publication-title: IEEE Trans. Neural Networks doi: 10.1109/72.279181 – volume: 8 start-page: 205 issue: 1 year: 1992 ident: 802_CR13 publication-title: Futur. Gener. Comput. Syst. doi: 10.1016/0167-739X(92)90040-I – year: 2023 ident: 802_CR1 publication-title: Information doi: 10.3390/info14020064 – ident: 802_CR5 doi: 10.1109/ISCA45697.2020.00016 – volume: 51 start-page: 1 issue: 1 year: 2022 ident: 802_CR9 publication-title: Int. J. Parallel Program. doi: 10.1007/s10766-022-00750-5
SSID	ssj0009788
Score	2.3580546
Snippet	Massive Data Parallel workloads, driven by inference on large ML models, are pushing hardware vendors to develop efficient and cost-effective multi-core server...
SourceID	proquest crossref springer
SourceType	Aggregation Database Index Database Publisher
StartPage	27
SubjectTerms	Artificial intelligence Central processing units Computer Science Computer vision CPUs Design Fashion models Hardware Inference Machine learning Neural networks Open standards Parallel processing Pipelining (computers) Processor Architectures Servers Software Software Engineering/Programming and Operating Systems Theory of Computation Workloads
Title	Analysis of Model Parallelism for AI Applications on a 64-core RV64 Server CPU
URI	https://link.springer.com/article/10.1007/s10766-025-00802-6 https://www.proquest.com/docview/3225632719
Volume	53
WOSCitedRecordID	wos001520151400001&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
journalDatabaseRights	– providerCode: PRVAVX databaseName: SpringerLINK Contemporary 1997-Present customDbUrl: eissn: 1573-7640 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0009788 issn: 0885-7458 databaseCode: RSV dateStart: 19970101 isFulltext: true titleUrlDefault: https://link.springer.com/search?facet-content-type=%22Journal%22 providerName: Springer Nature
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV1LS8NAEF60evBifWK1yh686UKSfR9LURSklGpLb2FfgUJNpan-fnfThFTRg56zGZLZeS2z830AXEfWeqsVDvl0ZxDhzCDpiESUulhRoqQtoZQmT3wwENOpHFZDYUV9271uSZaRemPYjbNwYZaickAUsW2w49OdCIQNo-dJA7XLS7ZJ7z4UcUJFNSrzs4yv6aipMb-1Rctsc9_-33cegP2quoS9tTkcgi2XH4F2zdwAK0c-BoMaiwQuMhjo0OZwqJaBVmU-K16hL2Rh7xH2NprbcJFDBRlBAfYSjiaMwBBmvND-cHwCxvd3L_0HVDErIJPwZIWswhn1By0qfLmqlZDEZtJxQSNiI66TTNvY-uIsMoIpijP_Q1w5rP1LCTZW4lPQyhe5OwMQK-cXCmUc48RlsVb-kEk0JkZrn_14B9zUCk7f1gAaaQOVHFSVelWlpapS1gHdeg_SypmKNMQchhMeyw64rXXePP5d2vnfll-AvaTctnC9rwtaq-W7uwS75mM1K5ZXpZF9AlBFyHo
linkProvider	Springer Nature
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV3dS8MwED90Cvri_MTp1Dz4poGuzUf7OIbD4SxjbmNvIW1SGMxO1unfb9K1dIo-6HOSo73cV7i73wHcOkoZqfU1Nu4uxoSzGAeaBJhS3ZKUyEDlUEqTPg9DfzoNBkVTWFZWu5cpydxSbzS7cWYLZinOG0Qx24YdYjyWRcwfvkwqqF2eT5s06kMxJ9QvWmV-pvHVHVUx5re0aO5tuvX_fechHBTRJWqvxeEItnR6DPVycgMqFPkEwhKLBC0SZMehzdFALu1Ylfkse0UmkEXtHmpvJLfRIkUSMYIt7CUaThhB1swYop3B-BTG3YdR5xEXkxVw7HJ3hZX0EmoeWtQ34Wok_YCoJNDcpw5RDo_cJFItZYIzJ_aZpF5ifohL7UXmkOvFKvDOoJYuUn0OyJPabPRlrBknOmlF0jwySeSROIqM9-MNuCsZLN7WABqigkq2rBKGVSJnlWANaJZ3IAplyoS1OcxzeStowH3J82r5d2oXf9t-A3uPo-e-6PfCp0vYd_MrtKV-Taitlu_6Cnbjj9UsW17nAvcJv8PLXg
linkToPdf	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LT8MwDLZgIMSF8RSDATlwg4g-8miP02BiYqoqYNNuVdqk0qTRTVvh95N0rToQHBDnplbi2LEt258Bri0ptdR6Cmtzl2DCWYJ9RXxMqbIFJcKXBZTSaMCDwBuP_XCti7-odq9SkqueBoPSlOV3c5nerTW-cWaKZykumkUx24QtYgrpTbz-Mqphd3kxeVKrEsWcUK9sm_mZxlfTVPub31KkheXpNf-_533YK71O1FmJyQFsqOwQmtVEB1Qq-BEEFUYJmqXIjEmbolAszLiV6WT5hrSDizp91FlLeqNZhgRiBBs4TPQ8YgSZ50cT7YbDYxj2Hl67j7icuIAThzs5lsJNqQ7AqKfd2Fh4PpGpr7hHLSItHjtpLG2pnTYr8ZigbqoPxIVyY_2T4ybSd0-gkc0ydQrIFUov9ESiGCcqtWOhg08SuySJY20VeQtuKmZH8xWwRlRDKBtWRZpVUcGqiLWgXd1HVCrZMjJvEXMdbvstuK34X3_-ndrZ35ZfwU5434sG_eDpHHad4gZNBWAbGvniXV3AdvKRT5aLy0L2PgFEj9RC
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%3Ajournal&rft.genre=article&rft.atitle=Analysis+of+Model+Parallelism+for+AI+Applications+on+a+64-core+RV64+Server+CPU&rft.jtitle=International+journal+of+parallel+programming&rft.au=Malenza%2C+Giulio&rft.au=Garcia%2C+Adriano+Marques&rft.au=Birke%2C+Robert&rft.au=Benini%2C+Luca&rft.date=2025-08-01&rft.pub=Springer+US&rft.issn=0885-7458&rft.eissn=1573-7640&rft.volume=53&rft.issue=4&rft_id=info:doi/10.1007%2Fs10766-025-00802-6&rft.externalDocID=10_1007_s10766_025_00802_6
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0885-7458&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0885-7458&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0885-7458&client=summon