{"id":"https://openalex.org/W7161745477","doi":"https://doi.org/10.48550/arxiv.2605.18404","title":"JanusPipe: Efficient Pipeline Parallel Training for Machine Learning Interatomic Potentials","display_name":"JanusPipe: Efficient Pipeline Parallel Training for Machine Learning Interatomic Potentials","publication_year":2026,"publication_date":"2026-05-18","ids":{"openalex":"https://openalex.org/W7161745477","doi":"https://doi.org/10.48550/arxiv.2605.18404"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.18404","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18404","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,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.18404","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136502296","display_name":"Hongyu Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Hongyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134051065","display_name":"Weijian Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Weijian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032135757","display_name":"Hongtao Xu","orcid":"https://orcid.org/0000-0001-5174-9079"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Hongtao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136491698","display_name":"Yan Wang","orcid":"https://orcid.org/0009-0000-0419-2087"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136464691","display_name":"Mingzhen Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Mingzhen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136478478","display_name":"Weile Jia","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jia, Weile","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136495278","display_name":"Guangming Tan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tan, Guangming","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":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11948","display_name":"Machine Learning in Materials Science","score":0.9315000176429749,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11948","display_name":"Machine Learning in Materials Science","score":0.9315000176429749,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11804","display_name":"Quantum many-body systems","score":0.009399999864399433,"subfield":{"id":"https://openalex.org/subfields/3107","display_name":"Atomic and Molecular Physics, and Optics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11471","display_name":"Block Copolymer Self-Assembly","score":0.008899999782443047,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.7768999934196472},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6722999811172485},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.5406000018119812},{"id":"https://openalex.org/keywords/parallelism","display_name":"Parallelism (grammar)","score":0.49779999256134033},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.42559999227523804},{"id":"https://openalex.org/keywords/interatomic-potential","display_name":"Interatomic potential","score":0.36309999227523804}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7894999980926514},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.7768999934196472},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6722999811172485},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.5406000018119812},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5055999755859375},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.49799999594688416},{"id":"https://openalex.org/C2781172179","wikidata":"https://www.wikidata.org/wiki/Q853109","display_name":"Parallelism (grammar)","level":2,"score":0.49779999256134033},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.42559999227523804},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4124000072479248},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.4000999927520752},{"id":"https://openalex.org/C2776372370","wikidata":"https://www.wikidata.org/wiki/Q3399989","display_name":"Interatomic potential","level":3,"score":0.36309999227523804},{"id":"https://openalex.org/C106515295","wikidata":"https://www.wikidata.org/wiki/Q26806595","display_name":"Parallel processing","level":2,"score":0.32600000500679016},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.32249999046325684},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3197000026702881},{"id":"https://openalex.org/C175309249","wikidata":"https://www.wikidata.org/wiki/Q725864","display_name":"Pipeline transport","level":2,"score":0.3027999997138977},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.28529998660087585},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.27390000224113464},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.2621000111103058},{"id":"https://openalex.org/C59593255","wikidata":"https://www.wikidata.org/wiki/Q901663","display_name":"Molecular dynamics","level":2,"score":0.2540000081062317}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.18404","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18404","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,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.18404","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18404","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,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Discovering":[0],"atom-level":[1],"phenomena":[2],"requires":[3],"molecular":[4],"dynamics":[5],"(MD)":[6],"simulations":[7],"with":[8,75],"ab":[9],"initio":[10],"accuracy.":[11],"Machine":[12],"learning":[13],"interatomic":[14],"potentials":[15],"(MLIPs)":[16],"enable":[17,102],"stable,":[18],"high-accuracy":[19],"MD":[20],"simulations,":[21],"and":[22,38,109,133,139],"their":[23],"models":[24],"exhibit":[25],"scaling-law":[26],"trends":[27],"similar":[28],"to":[29,49,101,111],"large":[30],"language":[31],"models.":[32],"However,":[33],"the":[34,67,117],"lack":[35],"of":[36],"scalable":[37],"efficient":[39,89],"distributed":[40,77],"training":[41,78,92],"systems":[42],"for":[43,81,95,106],"conservative":[44,54,96,107],"MLIPs":[45,55],"makes":[46],"them":[47],"difficult":[48],"scale.":[50],"This":[51,70],"is":[52],"because":[53],"inherently":[56],"follow":[57],"a":[58,73],"double-backward":[59],"execution":[60],"pattern,":[61],"which":[62],"involves":[63],"computing":[64],"gradients":[65],"during":[66],"forward":[68],"pass.":[69],"pattern":[71],"creates":[72],"mismatch":[74],"existing":[76],"systems,":[79],"especially":[80],"pipeline":[82,104,113],"parallelism.":[83],"Therefore,":[84],"we":[85],"present":[86],"JanusPipe,":[87],"an":[88],"3D-parallel":[90],"(PP/DP/GP)":[91],"system":[93],"tailored":[94],"MLIPs.":[97],"It":[98],"integrates":[99],"SymFold":[100],"memory-efficient":[103],"parallelism":[105],"MLIPs,":[108],"WaveK":[110],"reduce":[112],"bubbles":[114],"by":[115,131],"balancing":[116],"four-phase":[118],"compute":[119],"time.":[120],"Experimental":[121],"results":[122],"on":[123,135],"32":[124],"GPUs":[125],"show":[126],"that":[127],"JanusPipe":[128],"improves":[129],"throughput":[130],"$1.51\\times$":[132],"$1.45\\times$":[134],"average":[136],"over":[137],"1F1B":[138],"Hanayo,":[140],"respectively.":[141]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-05-20T00:00:00"}