{"id":"https://openalex.org/W7163526211","doi":"https://doi.org/10.48550/arxiv.2606.04847","title":"MusaCoder: Native GPU Kernel Generation with Full-Stack Training on Moore Threads GPU","display_name":"MusaCoder: Native GPU Kernel Generation with Full-Stack Training on Moore Threads GPU","publication_year":2026,"publication_date":"2026-06-03","ids":{"openalex":"https://openalex.org/W7163526211","doi":"https://doi.org/10.48550/arxiv.2606.04847"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.04847","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.04847","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"license_id":null,"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.2606.04847","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137914501","display_name":"Kun Cheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cheng, Kun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137811882","display_name":"Songshuo Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Songshuo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029274790","display_name":"Sicong Liao","orcid":"https://orcid.org/0000-0001-7340-520X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liao, Sicong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137815294","display_name":"Tankun Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Tankun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137909662","display_name":"Yafei Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yafei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137870911","display_name":"Dong Ho Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Dong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137809218","display_name":"Qiheng Lv","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lv, Qiheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137823268","display_name":"Hua Wang (13530)","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Hua","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137829298","display_name":"Zhi Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Zhi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5137877319","display_name":"Yaohua Tang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tang, Yaohua","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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.3483999967575073,"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"}},"topics":[{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.3483999967575073,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.22920000553131104,"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/T11948","display_name":"Machine Learning in Materials Science","score":0.06960000097751617,"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/kernel","display_name":"Kernel (algebra)","score":0.7408999800682068},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.6635000109672546},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5669000148773193},{"id":"https://openalex.org/keywords/correctness","display_name":"Correctness","score":0.5378999710083008},{"id":"https://openalex.org/keywords/thread","display_name":"Thread (computing)","score":0.44190001487731934},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.42910000681877136},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.40720000863075256},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.3691999912261963}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8561000227928162},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.7408999800682068},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.6635000109672546},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.569599986076355},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5669000148773193},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.5378999710083008},{"id":"https://openalex.org/C138101251","wikidata":"https://www.wikidata.org/wiki/Q213092","display_name":"Thread (computing)","level":2,"score":0.44190001487731934},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.42910000681877136},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.40720000863075256},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.40529999136924744},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.3691999912261963},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.3472000062465668},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3456999957561493},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3084000051021576},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.2957000136375427},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.28999999165534973},{"id":"https://openalex.org/C150552126","wikidata":"https://www.wikidata.org/wiki/Q339387","display_name":"SIMD","level":2,"score":0.2863999903202057},{"id":"https://openalex.org/C122280245","wikidata":"https://www.wikidata.org/wiki/Q620622","display_name":"Kernel method","level":3,"score":0.28299999237060547},{"id":"https://openalex.org/C202491316","wikidata":"https://www.wikidata.org/wiki/Q272683","display_name":"Instruction set","level":2,"score":0.2759000062942505},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.26100000739097595},{"id":"https://openalex.org/C201410400","wikidata":"https://www.wikidata.org/wiki/Q1064412","display_name":"Multithreading","level":3,"score":0.2551000118255615}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.04847","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.04847","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.04847","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.04847","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Native":[0],"GPU":[1,44],"kernel":[2,45,156],"generation":[3,46],"turns":[4],"high-level":[5],"tensor":[6],"programs":[7],"into":[8],"executable,":[9],"efficient":[10],"low-level":[11],"code.":[12],"Existing":[13],"Large":[14],"Language":[15],"Models":[16],"(LLMs)":[17],"struggle":[18],"with":[19,122],"this":[20],"task,":[21],"while":[22],"execution-based":[23],"reinforcement":[24],"learning":[25],"suffers":[26],"from":[27,90],"sparse":[28],"rewards,":[29,83],"reward":[30,72],"hacking,":[31],"and":[32,49,61,71,94,103,113,119,132,180],"training":[33,40,153,179],"instability.":[34],"We":[35],"present":[36],"MusaCoder,":[37],"a":[38,68,104,137,174],"full-stack":[39,151],"framework":[41],"for":[42,80,87,96,154,177],"native":[43,155],"on":[47,101,182],"CUDA":[48],"MUSA":[50],"backends.":[51],"MusaCoder":[52,77,109],"combines":[53],"progressive":[54],"kernel-oriented":[55],"data":[56],"synthesis,":[57],"diversity-preserving":[58],"rejection":[59],"fine-tuning,":[60],"execution-feedback":[62,152],"Reinforcement":[63],"Learning":[64],"(RL)":[65],"through":[66],"MooreEval,":[67],"distributed":[69],"verifier":[70],"environment.":[73],"To":[74],"stabilize":[75],"RL,":[76],"introduces":[78],"PrimeEcho":[79],"first-turn-anchored":[81],"multi-turn":[82],"Buffered":[84],"Dynamic":[85],"Retry":[86],"recovering":[88],"signals":[89],"all-failed":[91],"hard":[92],"samples,":[93],"MirrorPop":[95],"off-policy":[97],"sequence":[98],"filtering.":[99],"Experiments":[100],"KernelBench":[102],"MUSA-ported":[105],"variant":[106],"show":[107],"that":[108],"outperforms":[110],"strong":[111],"open-source":[112],"proprietary":[114],"baselines":[115],"in":[116],"both":[117],"correctness":[118],"empirical":[120],"speedup,":[121],"the":[123,133,141,148,160,168],"9B":[124],"model":[125,135],"matching":[126],"or":[127],"exceeding":[128],"frontier":[129],"closed-source":[130],"models":[131],"27B":[134],"establishing":[136],"new":[138],"state":[139],"of":[140,150,162],"art.":[142],"These":[143],"results":[144],"demonstrate":[145],"not":[146],"only":[147],"effectiveness":[149],"generation,":[157],"but":[158],"also":[159],"capability":[161],"Moore":[163],"Threads":[164],"GPUs":[165],"to":[166],"support":[167],"complete":[169],"LLM":[170],"post-training":[171],"stack,":[172],"providing":[173],"practical":[175],"foundation":[176],"large-model":[178],"optimization":[181],"emerging":[183],"accelerators.":[184]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-05T00:00:00"}