{"id":"https://openalex.org/W4412610767","doi":"https://doi.org/10.1109/ipdps64566.2025.00095","title":"Accelerating Sparse Linear Solvers on Intelligence Processing Units","display_name":"Accelerating Sparse Linear Solvers on Intelligence Processing Units","publication_year":2025,"publication_date":"2025-06-03","ids":{"openalex":"https://openalex.org/W4412610767","doi":"https://doi.org/10.1109/ipdps64566.2025.00095"},"language":"en","primary_location":{"id":"doi:10.1109/ipdps64566.2025.00095","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ipdps64566.2025.00095","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE International Parallel and Distributed Processing Symposium (IPDPS)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5119057139","display_name":"Tim Noack","orcid":null},"institutions":[{"id":"https://openalex.org/I31512782","display_name":"Technische Universit\u00e4t Darmstadt","ror":"https://ror.org/05n911h24","country_code":"DE","type":"education","lineage":["https://openalex.org/I31512782"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Tim Noack","raw_affiliation_strings":["TU Darmstadt,Embedded Systems and Applications,Darmstadt,Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"TU Darmstadt,Embedded Systems and Applications,Darmstadt,Germany","institution_ids":["https://openalex.org/I31512782"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Louis Kr\u00fcger","orcid":null},"institutions":[{"id":"https://openalex.org/I31512782","display_name":"Technische Universit\u00e4t Darmstadt","ror":"https://ror.org/05n911h24","country_code":"DE","type":"education","lineage":["https://openalex.org/I31512782"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Louis Kr\u00fcger","raw_affiliation_strings":["TU Darmstadt,Fluid Mechanics and Aerodynamics,Darmstadt,Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"TU Darmstadt,Fluid Mechanics and Aerodynamics,Darmstadt,Germany","institution_ids":["https://openalex.org/I31512782"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101408468","display_name":"Andreas Koch","orcid":"https://orcid.org/0000-0002-1044-3142"},"institutions":[{"id":"https://openalex.org/I31512782","display_name":"Technische Universit\u00e4t Darmstadt","ror":"https://ror.org/05n911h24","country_code":"DE","type":"education","lineage":["https://openalex.org/I31512782"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Andreas Koch","raw_affiliation_strings":["TU Darmstadt,Embedded Systems and Applications,Darmstadt,Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"TU Darmstadt,Embedded Systems and Applications,Darmstadt,Germany","institution_ids":["https://openalex.org/I31512782"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I31512782"],"apc_list":null,"apc_paid":null,"fwci":1.673,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.82973021,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"1023","last_page":"1035"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.8406000137329102,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"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/T10320","display_name":"Neural Networks and Applications","score":0.8406000137329102,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"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/T11697","display_name":"Numerical Methods and Algorithms","score":0.7925000190734863,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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":0.7432000041007996,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7152548432350159},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3448488414287567}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7152548432350159},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3448488414287567}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/ipdps64566.2025.00095","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ipdps64566.2025.00095","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE International Parallel and Distributed Processing Symposium (IPDPS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G8972483603","display_name":null,"funder_award_id":"01IS22091","funder_id":"https://openalex.org/F4320321114","funder_display_name":"Bundesministerium f\u00fcr Bildung und Forschung"}],"funders":[{"id":"https://openalex.org/F4320321114","display_name":"Bundesministerium f\u00fcr Bildung und Forschung","ror":"https://ror.org/04pz7b180"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":28,"referenced_works":["https://openalex.org/W71145039","https://openalex.org/W1506342804","https://openalex.org/W1520511539","https://openalex.org/W1845641575","https://openalex.org/W1978615272","https://openalex.org/W1988201922","https://openalex.org/W1998725414","https://openalex.org/W2020804487","https://openalex.org/W2035080386","https://openalex.org/W2039417226","https://openalex.org/W2042879774","https://openalex.org/W2053486661","https://openalex.org/W2080090223","https://openalex.org/W2111593426","https://openalex.org/W2128853364","https://openalex.org/W2155216327","https://openalex.org/W2482799402","https://openalex.org/W3043116807","https://openalex.org/W3165066978","https://openalex.org/W3166663340","https://openalex.org/W4281657881","https://openalex.org/W4316084067","https://openalex.org/W4361281982","https://openalex.org/W4388666543","https://openalex.org/W6788966007","https://openalex.org/W6793641609","https://openalex.org/W6852520788","https://openalex.org/W6857120838"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"This":[0],"paper":[1],"presents":[2],"an":[3,32],"open-source":[4],"framework":[5,62,114,209],"for":[6,148,222,240],"solving":[7,149],"large,":[8,150],"sparse":[9,151,242],"linear":[10,120,243],"systems":[11],"on":[12,183,199],"Intelligence":[13],"Processing":[14],"Units":[15],"(IPUs).":[16],"The":[17,82,138],"IPU":[18],"is":[19],"a":[20,96,116,146,228],"massively":[21],"parallel":[22,119],"architecture":[23],"consisting":[24],"of":[25,27,118,141,160,238],"thousands":[26],"independent":[28],"cores":[29],"connected":[30],"by":[31,78,216],"all-to-all":[33,110],"communication":[34,111],"fabric.":[35,112],"We":[36,94,153,178],"introduce":[37],"two":[38],"Domain-Specific":[39],"Languages,":[40],"CodeDSL":[41],"and":[42,54,74,122,132,167,189,196,213,224,248],"TensorDSL,":[43],"that":[44,101,207],"allow":[45],"us":[46,87],"to":[47,57,67,88,170],"express":[48],"complex":[49],"algebraic":[50],"algorithms":[51],"like":[52],"solvers":[53,121,226],"preconditioners":[55],"close":[56],"their":[58],"mathematical":[59],"notation.":[60],"Our":[61,113,233],"symbolically":[63],"executes":[64],"these":[65],"DSLs":[66],"generate":[68],"the":[69,75,108,125,133,157,161,186,236],"dataflow":[70],"graph,":[71],"execution":[72],"schedule,":[73],"codelets":[76],"required":[77],"IPU's":[79,83,109,139],"programming":[80],"model.":[81],"cacheless":[84],"design":[85],"allows":[86],"reorder":[89],"matrices":[90,184],"without":[91,174],"performance":[92,182],"penalties.":[93],"present":[95],"novel":[97,158],"matrix":[98],"reordering":[99],"strategy":[100],"enables":[102],"efficient":[103],"blockwise":[104],"halo":[105],"exchanges":[106],"with":[107],"implements":[115],"suite":[117],"preconditioners,":[123],"including":[124],"Preconditioned":[126],"Bi-Conjugate":[127],"Gradient":[128],"Stabilized":[129],"(PBiCGStab)":[130],"method":[131,166],"Incomplete":[134],"LU":[135],"Factorization":[136],"(ILU).":[137],"lack":[140],"native":[142,175],"double-precision":[143,176],"support":[144],"poses":[145],"challenge":[147],"systems.":[152],"address":[154],"this":[155],"using":[156],"combination":[159],"MixedPrecision":[162],"Iterative":[163],"Refinement":[164],"(MPIR)":[165],"double-word":[168],"arithmetics":[169],"achieve":[171],"high-precision":[172],"solutions":[173],"support.":[177],"evaluate":[179],"our":[180,208],"framework's":[181],"from":[185],"SuiteSparse":[187],"collection":[188],"synthetic":[190],"Poisson":[191],"problems,":[192],"demonstrating":[193],"near-ideal":[194],"strong":[195],"weak":[197],"scaling":[198],"Sparse":[200],"Matrix-Vector":[201],"Multiplications":[202],"(SpMVs).":[203],"Comparative":[204],"benchmarks":[205],"show":[206],"outperforms":[210],"state-of-the-art":[211],"CPU":[212],"GPU":[214],"implementations":[215],"up":[217],"to$150":[218],"x$and$36":[219],"x$,":[220],"respectively,":[221],"SpMVs":[223],"iterative":[225],"at":[227],"comparable":[229],"energy":[230],"consumption":[231],"level.":[232],"results":[234],"highlight":[235],"potential":[237],"IPUs":[239],"high-performance":[241],"algebra":[244],"computations":[245],"in":[246],"scientific":[247],"engineering":[249],"applications.":[250]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}