{"id":"https://openalex.org/W7133484784","doi":"https://doi.org/10.1109/hpca68181.2026.11408457","title":"WATOS: Efficient LLM Training Strategies and Architecture Co-Exploration for Wafer-Scale Chip","display_name":"WATOS: Efficient LLM Training Strategies and Architecture Co-Exploration for Wafer-Scale Chip","publication_year":2026,"publication_date":"2026-01-31","ids":{"openalex":"https://openalex.org/W7133484784","doi":"https://doi.org/10.1109/hpca68181.2026.11408457"},"language":null,"primary_location":{"id":"doi:10.1109/hpca68181.2026.11408457","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpca68181.2026.11408457","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5089405414","display_name":"Huizheng Wang","orcid":"https://orcid.org/0000-0002-9763-8208"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Huizheng Wang","raw_affiliation_strings":["School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084"],"affiliations":[{"raw_affiliation_string":"School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121146620","display_name":"Zichuan Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zichuan Wang","raw_affiliation_strings":["School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084"],"affiliations":[{"raw_affiliation_string":"School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100332091","display_name":"Hongbin Wang","orcid":"https://orcid.org/0000-0002-9828-9054"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hongbin Wang","raw_affiliation_strings":["School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084"],"affiliations":[{"raw_affiliation_string":"School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5084637087","display_name":"Jingxiang Hou","orcid":null},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jingxiang Hou","raw_affiliation_strings":["School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084"],"affiliations":[{"raw_affiliation_string":"School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069838515","display_name":"Tingcha Wei","orcid":"https://orcid.org/0000-0001-7619-4313"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Taiquan Wei","raw_affiliation_strings":["School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084"],"affiliations":[{"raw_affiliation_string":"School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128044651","display_name":"Chao Li","orcid":null},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chao Li","raw_affiliation_strings":["School of Computer Science and Engineering, Shanghai Jiao Tong University,Shanghai,China,200240"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Shanghai Jiao Tong University,Shanghai,China,200240","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068155662","display_name":"Yuanming Hu","orcid":"https://orcid.org/0000-0002-1136-9909"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang Hu","raw_affiliation_strings":["School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084"],"affiliations":[{"raw_affiliation_string":"School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5078843314","display_name":"Shouyi Yin","orcid":"https://orcid.org/0000-0002-8438-8588"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shouyi Yin","raw_affiliation_strings":["School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084"],"affiliations":[{"raw_affiliation_string":"School of Integrated Circuits, BNRist Tsinghua University,Beijing,China,100084","institution_ids":["https://openalex.org/I99065089"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5089405414"],"corresponding_institution_ids":["https://openalex.org/I99065089"],"apc_list":null,"apc_paid":null,"fwci":44.2394,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.99753509,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":97,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"19"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11527","display_name":"3D IC and TSV technologies","score":0.28850001096725464,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11527","display_name":"3D IC and TSV technologies","score":0.28850001096725464,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11338","display_name":"Advancements in Photolithography Techniques","score":0.14579999446868896,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10624","display_name":"Silicon and Solar Cell Technologies","score":0.08169999718666077,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.5353000164031982},{"id":"https://openalex.org/keywords/chip","display_name":"Chip","score":0.412200003862381},{"id":"https://openalex.org/keywords/architecture","display_name":"Architecture","score":0.38769999146461487},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.3084999918937683},{"id":"https://openalex.org/keywords/power","display_name":"Power (physics)","score":0.2863999903202057},{"id":"https://openalex.org/keywords/systems-architecture","display_name":"Systems architecture","score":0.28209999203681946}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5914999842643738},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.5353000164031982},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.41830000281333923},{"id":"https://openalex.org/C165005293","wikidata":"https://www.wikidata.org/wiki/Q1074500","display_name":"Chip","level":2,"score":0.412200003862381},{"id":"https://openalex.org/C123657996","wikidata":"https://www.wikidata.org/wiki/Q12271","display_name":"Architecture","level":2,"score":0.38769999146461487},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3084999918937683},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.30329999327659607},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.2915000021457672},{"id":"https://openalex.org/C163258240","wikidata":"https://www.wikidata.org/wiki/Q25342","display_name":"Power (physics)","level":2,"score":0.2863999903202057},{"id":"https://openalex.org/C98025372","wikidata":"https://www.wikidata.org/wiki/Q477538","display_name":"Systems architecture","level":3,"score":0.28209999203681946},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.27219998836517334},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.27079999446868896},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2669999897480011},{"id":"https://openalex.org/C118021083","wikidata":"https://www.wikidata.org/wiki/Q610398","display_name":"System on a chip","level":2,"score":0.26489999890327454},{"id":"https://openalex.org/C201995342","wikidata":"https://www.wikidata.org/wiki/Q682496","display_name":"Systems engineering","level":1,"score":0.26440000534057617},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C147494362","wikidata":"https://www.wikidata.org/wiki/Q2078905","display_name":"Troubleshooting","level":2,"score":0.251800000667572}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpca68181.2026.11408457","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpca68181.2026.11408457","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.4625702500343323,"display_name":"Industry, innovation and infrastructure","id":"https://metadata.un.org/sdg/9"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":120,"referenced_works":["https://openalex.org/W1586409345","https://openalex.org/W1975292844","https://openalex.org/W2025516544","https://openalex.org/W2087784837","https://openalex.org/W2119391823","https://openalex.org/W2123016589","https://openalex.org/W2124653173","https://openalex.org/W2131613942","https://openalex.org/W2142246791","https://openalex.org/W2289252105","https://openalex.org/W2753840902","https://openalex.org/W2794478957","https://openalex.org/W2883830791","https://openalex.org/W2886064645","https://openalex.org/W2887063112","https://openalex.org/W2913573286","https://openalex.org/W2929862812","https://openalex.org/W2935331687","https://openalex.org/W2940862705","https://openalex.org/W2980104813","https://openalex.org/W2980200167","https://openalex.org/W2981849318","https://openalex.org/W2991040477","https://openalex.org/W3016212306","https://openalex.org/W3032324753","https://openalex.org/W3081168214","https://openalex.org/W3102790199","https://openalex.org/W3118922386","https://openalex.org/W3129831491","https://openalex.org/W3132107458","https://openalex.org/W3135160281","https://openalex.org/W3135807226","https://openalex.org/W3179227704","https://openalex.org/W3182615510","https://openalex.org/W3187788856","https://openalex.org/W3190806564","https://openalex.org/W3192636176","https://openalex.org/W3200652633","https://openalex.org/W3204998121","https://openalex.org/W3205803342","https://openalex.org/W3206832494","https://openalex.org/W3210059290","https://openalex.org/W3211730428","https://openalex.org/W4223941905","https://openalex.org/W4231702991","https://openalex.org/W4253012315","https://openalex.org/W4280524580","https://openalex.org/W4280633094","https://openalex.org/W4281790033","https://openalex.org/W4283643682","https://openalex.org/W4288083528","https://openalex.org/W4293024081","https://openalex.org/W4297097318","https://openalex.org/W4297097426","https://openalex.org/W4308083753","https://openalex.org/W4313590988","https://openalex.org/W4319988867","https://openalex.org/W4327930463","https://openalex.org/W4360831791","https://openalex.org/W4360831999","https://openalex.org/W4380881063","https://openalex.org/W4381886086","https://openalex.org/W4385525225","https://openalex.org/W4385532514","https://openalex.org/W4385562710","https://openalex.org/W4386348101","https://openalex.org/W4386709668","https://openalex.org/W4386768656","https://openalex.org/W4387934966","https://openalex.org/W4388031315","https://openalex.org/W4388697601","https://openalex.org/W4389459396","https://openalex.org/W4389476193","https://openalex.org/W4389491903","https://openalex.org/W4389491911","https://openalex.org/W4390415044","https://openalex.org/W4392567268","https://openalex.org/W4393406935","https://openalex.org/W4393407116","https://openalex.org/W4393407138","https://openalex.org/W4393592481","https://openalex.org/W4393851766","https://openalex.org/W4394627396","https://openalex.org/W4395106467","https://openalex.org/W4395117922","https://openalex.org/W4400033923","https://openalex.org/W4400322767","https://openalex.org/W4400680190","https://openalex.org/W4400681611","https://openalex.org/W4401211642","https://openalex.org/W4401211750","https://openalex.org/W4401212149","https://openalex.org/W4401212157","https://openalex.org/W4402475781","https://openalex.org/W4402570070","https://openalex.org/W4404037868","https://openalex.org/W4404954358","https://openalex.org/W4404955085","https://openalex.org/W4405220108","https://openalex.org/W4405754937","https://openalex.org/W4406166720","https://openalex.org/W4407129663","https://openalex.org/W4407197043","https://openalex.org/W4408174495","https://openalex.org/W4408183166","https://openalex.org/W4408183245","https://openalex.org/W4409362467","https://openalex.org/W4411485931","https://openalex.org/W4411486243","https://openalex.org/W4411486513","https://openalex.org/W4411486567","https://openalex.org/W4411688694","https://openalex.org/W4413017445","https://openalex.org/W4414197279","https://openalex.org/W4414198924","https://openalex.org/W4415293446","https://openalex.org/W4415792806","https://openalex.org/W4415792849","https://openalex.org/W7092191477","https://openalex.org/W7092205305"],"related_works":[],"abstract_inverted_index":{"Training":[0],"large":[1],"language":[2],"models":[3],"(LLMs)":[4],"imposes":[5,75],"extreme":[6],"demands":[7],"on":[8,131,135],"computation,":[9],"memory":[10,159],"capacity,":[11],"and":[12,21,52,83,110,140,152,174,194],"interconnect":[13],"bandwidth,":[14],"driven":[15],"by":[16,31],"their":[17],"ever-increasing":[18],"parameter":[19],"scales":[20],"intensive":[22],"data":[23],"movement.":[24],"Wafer-scale":[25],"integration":[26,61],"offers":[27],"a":[28,103,116],"promising":[29],"solution":[30],"densely":[32],"integrating":[33],"multiple":[34],"single-die":[35],"chips":[36,147],"with":[37,217],"high-speed":[38],"die-to-die":[39],"(D2D)":[40],"interconnects.":[41],"However,":[42],"the":[43,57,79,98,136,158,167,218],"limited":[44],"wafer":[45,178],"area":[46],"necessitates":[47],"trade-offs":[48],"among":[49],"compute,":[50],"memory,":[51],"communication":[53],"resources.":[54],"Fully":[55],"harnessing":[56],"potential":[58],"of":[59,81,189,220],"wafer-scale":[60,111,146,214],"while":[62],"mitigating":[63],"its":[64],"architectural":[65,125],"constraints":[66],"is":[67],"essential":[68],"for":[69,78,106,127],"maximizing":[70],"LLM":[71,107,163,170,201,221],"training":[72,84,108,171,179,219],"performance.":[73],"This":[74],"significant":[76],"challenges":[77],"co-optimization":[80],"architecture":[82,215],"strategies.":[85],"Unfortunately,":[86],"existing":[87],"approaches":[88],"all":[89],"fall":[90],"short":[91],"in":[92],"addressing":[93,157],"these":[94],"challenges.":[95],"To":[96],"bridge":[97],"gap,":[99],"we":[100,133,206],"propose":[101],"WATOS,":[102],"co-exploration":[104],"framework":[105,172],"strategy":[109],"architecture.":[112],"We":[113],"first":[114],"define":[115],"highly":[117],"configurable":[118],"hardware":[119],"template":[120],"designed":[121],"to":[122,145,148,166,209],"explore":[123,149],"optimal":[124,150],"parameters":[126],"waferscale":[128],"chips.":[129],"Based":[130],"it,":[132],"capitalize":[134],"high":[137],"D2D":[138],"bandwidth":[139],"fine-grained":[141],"operation":[142],"advantages":[143],"inherent":[144],"parallelism":[151],"resource":[153],"allocation":[154],"strategies,":[155],"effectively":[156],"underutilization":[160],"issues":[161],"during":[162],"training.":[164],"Compared":[165],"state-of-the-art":[168],"(SOTA)":[169],"Megatron":[173],"Cerebras'":[175],"weight":[176],"streaming":[177],"strategy,":[180],"WATOS":[181,208],"can":[182],"achieve":[183],"an":[184],"average":[185],"overall":[186],"throughput":[187],"improvement":[188],"<tex":[190,195],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[191,196],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$2.74":[192],"\\times$</tex>":[193,198],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$1.53":[197],"across":[199],"various":[200],"models,":[202],"respectively.":[203],"In":[204],"addition,":[205],"leverage":[207],"reveal":[210],"intriguing":[211],"insights":[212],"about":[213],"design":[216],"workloads.":[222]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-03-06T06:45:51.903784","created_date":"2026-03-05T00:00:00"}
