{"id":"https://openalex.org/W4409132068","doi":"https://doi.org/10.1109/hpec62836.2024.10938418","title":"Breakthrough Low-Latency, High-Energy-Efficiency LLM Inference Performance Using NorthPole","display_name":"Breakthrough Low-Latency, High-Energy-Efficiency LLM Inference Performance Using NorthPole","publication_year":2024,"publication_date":"2024-09-23","ids":{"openalex":"https://openalex.org/W4409132068","doi":"https://doi.org/10.1109/hpec62836.2024.10938418"},"language":"en","primary_location":{"id":"doi:10.1109/hpec62836.2024.10938418","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec62836.2024.10938418","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE High Performance Extreme Computing Conference (HPEC)","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/A5022262317","display_name":"Rathinakumar Appuswamy","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Rathinakumar Appuswamy","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027380988","display_name":"Michael DeBole","orcid":"https://orcid.org/0000-0002-3068-9430"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Michael V. Debole","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041605470","display_name":"Brian Taba","orcid":"https://orcid.org/0009-0008-7831-1139"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Brian Taba","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5055044754","display_name":"Steven K. Esser","orcid":"https://orcid.org/0009-0008-4089-0577"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Steven K. Esser","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5044024456","display_name":"Andrew S. Cassidy","orcid":"https://orcid.org/0000-0001-7305-4198"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Andrew S. Cassidy","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101757788","display_name":"Arnon Amir","orcid":"https://orcid.org/0009-0000-6814-165X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Arnon Amir","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093091740","display_name":"Alexander Andreopoulos","orcid":"https://orcid.org/0009-0001-2408-5171"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Alexander Andreopoulos","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5045970662","display_name":"Deepika Bablani","orcid":"https://orcid.org/0000-0002-8987-8201"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Deepika Bablani","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064191197","display_name":"Pallab Datta","orcid":"https://orcid.org/0000-0002-0515-9729"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pallab Datta","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069592415","display_name":"Jeffrey A. Kusnitz","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jeffrey A. Kusnitz","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5059841005","display_name":"Nathaniel J. McClatchey","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nathaniel J. McClatchey","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5089195359","display_name":"Neil McGlohon","orcid":"https://orcid.org/0000-0001-8304-4911"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Neil McGlohon","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5059206462","display_name":"Jeffrey L. McKinstry","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jeffrey L. McKinstry","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102976698","display_name":"Tapan K. Nayak","orcid":"https://orcid.org/0000-0001-8941-8961"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tapan K. Nayak","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111990430","display_name":"Daniel F. Smith","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Daniel F. Smith","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5056962045","display_name":"Rafael T. de Sousa","orcid":"https://orcid.org/0000-0003-1101-3029"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rafael Sousa","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054567571","display_name":"Ignacio Terrizzano","orcid":"https://orcid.org/0000-0001-6179-2601"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ignacio Terrizzano","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002319629","display_name":"Filipp Akopyan","orcid":"https://orcid.org/0009-0001-5312-1311"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Filipp Akopyan","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031497317","display_name":"Peter J. Carlson","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peter J. Carlson","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092947158","display_name":"Rajamohan Gandhasri","orcid":"https://orcid.org/0009-0003-5068-1924"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rajamohan Gandhasri","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005670873","display_name":"Guillaume Garreau","orcid":"https://orcid.org/0000-0002-4115-9232"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guillaume J. Garreau","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102979275","display_name":"Nelson Mimura Gonzalez","orcid":"https://orcid.org/0000-0002-8313-8114"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nelson M. Gonzalez","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042484827","display_name":"Megumi Ito","orcid":"https://orcid.org/0009-0002-3627-5308"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Megumi Ito","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5084934945","display_name":"Jennifer L. Klamo","orcid":"https://orcid.org/0009-0001-9042-0476"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jennifer L. Klamo","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5063752378","display_name":"Yutaka Nakamura","orcid":"https://orcid.org/0000-0002-6921-1266"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yutaka Nakamura","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069172065","display_name":"Carlos Ortega Otero","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Carlos Ortega Otero","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042044567","display_name":"W. P. Risk","orcid":"https://orcid.org/0009-0004-3211-0738"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"William P. Risk","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5083548741","display_name":"Jun Sawada","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jun Sawada","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113582614","display_name":"Kai Schleupen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kai Schleupen","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092947159","display_name":"Jay Sivagnaname","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jay Sivagnaname","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093875966","display_name":"Matthew Stallone","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Matthew Stallone","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5011340584","display_name":"Takanori Ueda","orcid":"https://orcid.org/0000-0003-1637-4640"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Takanori Ueda","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005956938","display_name":"Myron Flickner","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Myron D. Flickner","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087078947","display_name":"John V. Arthur","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"John V. Arthur","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5049734237","display_name":"Rameswar Panda","orcid":"https://orcid.org/0000-0003-4359-2475"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rameswar Panda","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100763986","display_name":"David Cox","orcid":"https://orcid.org/0000-0002-2189-9743"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"David D. Cox","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5063718511","display_name":"Dharmendra S. Modha","orcid":"https://orcid.org/0009-0005-6714-8127"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dharmendra S. Modha","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":37,"corresponding_author_ids":["https://openalex.org/A5022262317"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.7866,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.75268686,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":96,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"8"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10688","display_name":"Image and Signal Denoising Methods","score":0.9664999842643738,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10688","display_name":"Image and Signal Denoising Methods","score":0.9664999842643738,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10876","display_name":"Fault Detection and Control Systems","score":0.9510999917984009,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems 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/T11447","display_name":"Blind Source Separation Techniques","score":0.9247000217437744,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6946258544921875},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.6299391388893127},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6031304001808167},{"id":"https://openalex.org/keywords/efficient-energy-use","display_name":"Efficient energy use","score":0.5544723272323608},{"id":"https://openalex.org/keywords/low-latency","display_name":"Low latency (capital markets)","score":0.4390266239643097},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.25625866651535034},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.1890236735343933},{"id":"https://openalex.org/keywords/telecommunications","display_name":"Telecommunications","score":0.1425361931324005},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.12296426296234131},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.1200236976146698}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6946258544921875},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.6299391388893127},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6031304001808167},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.5544723272323608},{"id":"https://openalex.org/C46637626","wikidata":"https://www.wikidata.org/wiki/Q6693015","display_name":"Low latency (capital markets)","level":2,"score":0.4390266239643097},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.25625866651535034},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.1890236735343933},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.1425361931324005},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.12296426296234131},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.1200236976146698}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpec62836.2024.10938418","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec62836.2024.10938418","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE High Performance Extreme Computing Conference (HPEC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Affordable and clean energy","score":0.8700000047683716,"id":"https://metadata.un.org/sdg/7"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":4,"referenced_works":["https://openalex.org/W1999563089","https://openalex.org/W4387772271","https://openalex.org/W4392746564","https://openalex.org/W6760069825"],"related_works":["https://openalex.org/W3128807919","https://openalex.org/W3176411177","https://openalex.org/W3205411230","https://openalex.org/W4286899009","https://openalex.org/W9168048","https://openalex.org/W4300849822","https://openalex.org/W4376480820","https://openalex.org/W3155891479","https://openalex.org/W3029351463","https://openalex.org/W4308600690"],"abstract_inverted_index":{"For":[0],"a":[1,4,16,42,48,65],"3-billion-parameter":[2],"LLM,":[3],"research":[5],"prototype":[6],"inference":[7],"appliance":[8],"with":[9],"16":[10,38],"IBM":[11],"AIU":[12],"NorthPole":[13,39,58,81],"processors":[14],"delivers":[15],"massive":[17],"28,356":[18],"tokens":[19],"/":[20,28],"second":[21],"of":[22,67],"system":[23],"throughput":[24],"and":[25,53],"sub-1":[26],"ms":[27],"token":[29],"(per-user)":[30],"latency":[31,52],"while":[32,88],"consuming":[33],"merely":[34],"672":[35],"W":[36],"for":[37],"cards":[40],"in":[41],"compact":[43],"2U":[44],"form":[45],"factor.":[46],"With":[47],"focus":[49],"on":[50],"low":[51],"high":[54],"energy":[55,85],"efficiency,":[56],"when":[57],"(in":[59,69],"12":[60],"nm)":[61,71],"is":[62],"compared":[63],"to":[64],"suite":[66],"GPUs":[68],"7/5/4":[70],"at":[72,76],"various":[73],"power":[74],"consumptions,":[75],"the":[77],"lowest":[78],"GPU":[79],"latency,":[80],"provides":[82],"72.7x":[83],"better":[84,90],"metric":[86],"(tokens/second/W)":[87],"providing":[89],"latency.":[91]},"counts_by_year":[{"year":2025,"cited_by_count":3}],"updated_date":"2026-03-04T07:04:00.330322","created_date":"2025-10-10T00:00:00"}