{"id":"https://openalex.org/W7207791791","doi":"https://doi.org/10.1109/hoti70696.2026.00020","title":"Scaling Inference Prefill with High-Radix Photonic Interconnects","display_name":"Scaling Inference Prefill with High-Radix Photonic Interconnects","publication_year":2026,"publication_date":"2026-08-01","ids":{"openalex":"https://openalex.org/W7207791791","doi":"https://doi.org/10.1109/hoti70696.2026.00020"},"language":"en","primary_location":{"id":"doi:10.1109/hoti70696.2026.00020","doi":"https://doi.org/10.1109/hoti70696.2026.00020","is_oa":true,"landing_page_url":"https://doi.org/10.1109/hoti70696.2026.00020","pdf_url":null,"source":{"id":"https://openalex.org/S4210232710","display_name":"IEEE Symposium on High-Performance Interconnects (HOTI)","issn_l":"1550-4794","issn":["1550-4794","2332-5569"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core"],"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"conference"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE Symposium on High-Performance Interconnects (HOTI)","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.1109/hoti70696.2026.00020","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5104561643","display_name":"Arulselvan Madhavan","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Arulselvan Madhavan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125890766","display_name":"Peter Carson","orcid":"https://orcid.org/0009-0006-4304-251X","observed_orcids":["https://orcid.org/0009-0006-4304-251X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peter Carson","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5147728332","display_name":"Taylor Groves","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Taylor Groves","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5149759704","display_name":"Thomas Graham","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Thomas Graham","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.0,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"14","last_page":"27"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10829","display_name":"Interconnection Networks and Systems","score":0.7627000212669373,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10829","display_name":"Interconnection Networks and Systems","score":0.7627000212669373,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.10949999839067459,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11527","display_name":"3D IC and TSV technologies","score":0.05130000039935112,"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/optical-interconnects","display_name":"optical interconnects","score":0.9340000152587891},{"id":"https://openalex.org/keywords/mixture-of-experts-models","display_name":"mixture-of-experts models","score":0.9100000262260437},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8610000014305115},{"id":"https://openalex.org/keywords/time-to-first-token","display_name":"time-to-first-token","score":0.7350000143051147},{"id":"https://openalex.org/keywords/agentic-ai","display_name":"agentic AI","score":0.671999990940094}],"concepts":[{"id":"https://openalex.org/C20788544","wikidata":"https://www.wikidata.org/wiki/Q467054","display_name":"Photonics","level":2,"score":0.7469000220298767},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7386999726295471},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6881999969482422},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.5916000008583069},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5404999852180481},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.48570001125335693},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.4699000120162964},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.4309000074863434},{"id":"https://openalex.org/C2778476105","wikidata":"https://www.wikidata.org/wiki/Q628539","display_name":"Workload","level":2,"score":0.37959998846054077},{"id":"https://openalex.org/C24326235","wikidata":"https://www.wikidata.org/wiki/Q126095","display_name":"Electronic engineering","level":1,"score":0.3407999873161316},{"id":"https://openalex.org/C22799297","wikidata":"https://www.wikidata.org/wiki/Q523846","display_name":"Photonic integrated circuit","level":3,"score":0.33899998664855957},{"id":"https://openalex.org/C183322885","wikidata":"https://www.wikidata.org/wiki/Q17007702","display_name":"Context model","level":3,"score":0.32280001044273376},{"id":"https://openalex.org/C53833338","wikidata":"https://www.wikidata.org/wiki/Q1061424","display_name":"Context switch","level":2,"score":0.31200000643730164},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.3082999885082245},{"id":"https://openalex.org/C132943942","wikidata":"https://www.wikidata.org/wiki/Q2562511","display_name":"Footprint","level":2,"score":0.2865000069141388},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.2851000130176544},{"id":"https://openalex.org/C123745756","wikidata":"https://www.wikidata.org/wiki/Q1665949","display_name":"Interconnection","level":2,"score":0.28450000286102295},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.28290000557899475},{"id":"https://openalex.org/C74912251","wikidata":"https://www.wikidata.org/wiki/Q6815727","display_name":"Memory footprint","level":2,"score":0.25200000405311584}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/hoti70696.2026.00020","doi":"https://doi.org/10.1109/hoti70696.2026.00020","is_oa":true,"landing_page_url":"https://doi.org/10.1109/hoti70696.2026.00020","pdf_url":null,"source":{"id":"https://openalex.org/S4210232710","display_name":"IEEE Symposium on High-Performance Interconnects (HOTI)","issn_l":"1550-4794","issn":["1550-4794","2332-5569"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core"],"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"conference"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE Symposium on High-Performance Interconnects (HOTI)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2609.01821","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.01821","pdf_url":"https://arxiv.org/pdf/2609.01821","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.01821","doi":"https://doi.org/10.48550/arxiv.2609.01821","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.01821","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.1109/hoti70696.2026.00020","doi":"https://doi.org/10.1109/hoti70696.2026.00020","is_oa":true,"landing_page_url":"https://doi.org/10.1109/hoti70696.2026.00020","pdf_url":null,"source":{"id":"https://openalex.org/S4210232710","display_name":"IEEE Symposium on High-Performance Interconnects (HOTI)","issn_l":"1550-4794","issn":["1550-4794","2332-5569"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core"],"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"conference"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE Symposium on High-Performance Interconnects (HOTI)","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.4943999946117401,"id":"https://metadata.un.org/sdg/9","display_name":"Industry, innovation and infrastructure"}],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7207791791.pdf","grobid_xml":"https://content.openalex.org/works/W7207791791.grobid-xml"},"referenced_works_count":8,"referenced_works":["https://openalex.org/W2912012512","https://openalex.org/W4280611847","https://openalex.org/W4391045969","https://openalex.org/W4405173907","https://openalex.org/W4405903187","https://openalex.org/W4415947799","https://openalex.org/W4416429967","https://openalex.org/W7104000765"],"related_works":["https://openalex.org/W4416429967","https://openalex.org/W4402835888","https://openalex.org/W7161719461","https://openalex.org/W4416023851","https://openalex.org/W7140347238","https://openalex.org/W4415249480","https://openalex.org/W4400495010","https://openalex.org/W4412084913","https://openalex.org/W4412084821","https://openalex.org/W7170161234"],"abstract_inverted_index":{"With":[0],"the":[1,10,20,33,55,106,121],"rise":[2],"of":[3,24,35],"inference":[4,40],"as":[5],"today\u2019s":[6],"dominant":[7],"AI":[8],"workload,":[9],"industry":[11],"is":[12],"transitioning":[13],"to":[14,18,125],"high-bandwidth":[15],"photonic":[16,37],"interconnects":[17,38],"meet":[19],"large":[21,56],"scale-up":[22],"requirements":[23],"increasingly":[25],"complex":[26],"Mixture-of-Experts":[27],"(MoE)":[28],"models.":[29],"This":[30],"paper":[31],"quantifies":[32],"benefits":[34],"3D-integrated":[36],"for":[39,48,61],"prefill":[41],"by":[42],"analyzing":[43],"tradeoffs":[44],"between":[45],"high-concurrency":[46],"throughput":[47],"Large":[49],"Language":[50],"Model":[51],"(LLM)":[52],"chat":[53],"and":[54,63,79,93,110],"context":[57,72,76,81],"windows":[58],"typically":[59],"required":[60,124],"reasoning":[62],"agentic":[64],"AI.":[65],"We":[66,84,100],"simulate":[67],"three":[68],"MoE":[69],"models:":[70],"short":[71],"(1K\u20138K":[73],"tokens),":[74,78],"medium":[75],"(128K":[77],"long":[80],"(1M":[82],"tokens).":[83],"evaluate":[85],"this":[86],"workload":[87],"across":[88,131],"existing":[89],"copper-based":[90],"GPU":[91],"systems":[92,136],"one":[94],"with":[95],"high":[96],"bandwidth":[97],"integrated":[98],"photonics.":[99],"show":[101],"2.1\u20133.2\u00d7":[102],"latency":[103],"improvements":[104,112],"in":[105,115],"stressed":[107],"high-batch":[108],"regimes":[109],"2.8\u20135.8\u00d7":[111],"over":[113],"baselines":[114],"communication-limited":[116],"configurations.":[117],"3D":[118],"photonics":[119],"enable":[120],"1152-GPU":[122],"footprint":[123],"lower":[126],"time-to-first-token,":[127],"yielding":[128],"2.2\u20134.5\u00d7":[129],"speedups":[130],"production-grade":[132],"platforms":[133],"when":[134],"electrical":[135],"cross":[137],"their":[138],"inherent":[139],"scale-up-pod":[140],"limits.":[141]},"counts_by_year":[],"updated_date":"2026-11-09T13:22:03.289127","created_date":"2026-09-04T00:00:00"}