{"id":"https://openalex.org/W7162194745","doi":"https://doi.org/10.1145/3816433","title":"CPU-GPU Workload Distribution during Throughput-Oriented LLM Inference on Single-GPU Systems","display_name":"CPU-GPU Workload Distribution during Throughput-Oriented LLM Inference on Single-GPU Systems","publication_year":2026,"publication_date":"2026-05-23","ids":{"openalex":"https://openalex.org/W7162194745","doi":"https://doi.org/10.1145/3816433"},"language":"en","primary_location":{"id":"doi:10.1145/3816433","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3816433","pdf_url":null,"source":{"id":"https://openalex.org/S26056741","display_name":"ACM Transactions on Architecture and Code Optimization","issn_l":"1544-3566","issn":["1544-3566","1544-3973"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Architecture and Code Optimization","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://doi.org/10.1145/3816433","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5051797786","display_name":"Daon Park","orcid":"https://orcid.org/0000-0003-2312-3049"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Daon Park","raw_affiliation_strings":["Seoul National University"],"raw_orcid":"https://orcid.org/0000-0003-2312-3049","affiliations":[{"raw_affiliation_string":"Seoul National University","institution_ids":["https://openalex.org/I139264467"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5136846826","display_name":"Bernhard Egger","orcid":"https://orcid.org/0000-0002-6645-6161"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]},{"id":"https://openalex.org/I81007117","display_name":"Lucerne University of Applied Sciences and Arts","ror":"https://ror.org/04nd0xd48","country_code":"CH","type":"education","lineage":["https://openalex.org/I81007117"]}],"countries":["CH","KR"],"is_corresponding":false,"raw_author_name":"Bernhard Egger","raw_affiliation_strings":["Lucerne University of Applied Sciences and Arts","School of Computer Science and Engineering, Seoul National University"],"raw_orcid":"https://orcid.org/0000-0002-6645-6161","affiliations":[{"raw_affiliation_string":"Lucerne University of Applied Sciences and Arts","institution_ids":["https://openalex.org/I81007117"]},{"raw_affiliation_string":"School of Computer Science and Engineering, Seoul National University","institution_ids":["https://openalex.org/I139264467"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.67791302,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"23","issue":"2","first_page":"1","last_page":"25"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T14347","display_name":"Big Data and Digital Economy","score":0.46219998598098755,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T14347","display_name":"Big Data and Digital Economy","score":0.46219998598098755,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10181","display_name":"Natural Language Processing Techniques","score":0.1315000057220459,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.060499999672174454,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/workload","display_name":"Workload","score":0.7598999738693237},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.6758999824523926},{"id":"https://openalex.org/keywords/idle","display_name":"Idle","score":0.6294999718666077},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5164999961853027},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5024999976158142},{"id":"https://openalex.org/keywords/hierarchy","display_name":"Hierarchy","score":0.4641000032424927},{"id":"https://openalex.org/keywords/state","display_name":"State (computer science)","score":0.4607999920845032},{"id":"https://openalex.org/keywords/memory-hierarchy","display_name":"Memory hierarchy","score":0.36160001158714294}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8647000193595886},{"id":"https://openalex.org/C2778476105","wikidata":"https://www.wikidata.org/wiki/Q628539","display_name":"Workload","level":2,"score":0.7598999738693237},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.6758999824523926},{"id":"https://openalex.org/C16320812","wikidata":"https://www.wikidata.org/wiki/Q1812200","display_name":"Idle","level":2,"score":0.6294999718666077},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5164999961853027},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5024999976158142},{"id":"https://openalex.org/C31170391","wikidata":"https://www.wikidata.org/wiki/Q188619","display_name":"Hierarchy","level":2,"score":0.4641000032424927},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.4607999920845032},{"id":"https://openalex.org/C2778100165","wikidata":"https://www.wikidata.org/wiki/Q1589327","display_name":"Memory hierarchy","level":3,"score":0.36160001158714294},{"id":"https://openalex.org/C2779439359","wikidata":"https://www.wikidata.org/wiki/Q317088","display_name":"Commodity","level":2,"score":0.3589000105857849},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.3538999855518341},{"id":"https://openalex.org/C2778348673","wikidata":"https://www.wikidata.org/wiki/Q739302","display_name":"Production (economics)","level":2,"score":0.3228999972343445},{"id":"https://openalex.org/C172658912","wikidata":"https://www.wikidata.org/wiki/Q661613","display_name":"Batch processing","level":2,"score":0.31839999556541443},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.2874000072479248},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2800999879837036},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.2761000096797943},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.27300000190734863},{"id":"https://openalex.org/C110121322","wikidata":"https://www.wikidata.org/wiki/Q865811","display_name":"Distribution (mathematics)","level":2,"score":0.26820001006126404},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.2572000026702881},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2535000145435333}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3816433","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3816433","pdf_url":null,"source":{"id":"https://openalex.org/S26056741","display_name":"ACM Transactions on Architecture and Code Optimization","issn_l":"1544-3566","issn":["1544-3566","1544-3973"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Architecture and Code Optimization","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1145/3816433","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3816433","pdf_url":null,"source":{"id":"https://openalex.org/S26056741","display_name":"ACM Transactions on Architecture and Code Optimization","issn_l":"1544-3566","issn":["1544-3566","1544-3973"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Architecture and Code Optimization","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G8532035263","display_name":null,"funder_award_id":"10077609","funder_id":"https://openalex.org/F4320321681","funder_display_name":"Ministry of Trade, Industry and Energy"}],"funders":[{"id":"https://openalex.org/F4320321681","display_name":"Ministry of Trade, Industry and Energy","ror":"https://ror.org/008nkqk13"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":20,"referenced_works":["https://openalex.org/W2888482885","https://openalex.org/W2896838216","https://openalex.org/W2964110616","https://openalex.org/W2984100107","https://openalex.org/W2998183051","https://openalex.org/W3131724164","https://openalex.org/W3133458480","https://openalex.org/W3165031368","https://openalex.org/W4220738659","https://openalex.org/W4321448364","https://openalex.org/W4321636575","https://openalex.org/W4378189609","https://openalex.org/W4385438314","https://openalex.org/W4387321091","https://openalex.org/W4389518760","https://openalex.org/W4395020691","https://openalex.org/W4402805837","https://openalex.org/W4403337152","https://openalex.org/W4403337153","https://openalex.org/W4404401018"],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2,118],"(LLMs)":[3],"have":[4,51,63],"recently":[5],"achieved":[6],"remarkable":[7],"performance":[8],"in":[9,26],"text":[10],"generation,":[11],"capturing":[12],"the":[13,23,31,65,73,86,92,129,132,147,150],"attention":[14],"of":[15,33,75,149,153],"a":[16,99],"broad":[17],"audience.":[18],"This":[19],"success,":[20],"driven":[21],"by":[22,71,91],"rapid":[24],"growth":[25],"model":[27,154],"parameters,":[28],"comes":[29],"at":[30],"expense":[32],"significantly":[34],"higher":[35,143],"operational":[36],"costs":[37],"and":[38,131,158],"decreased":[39],"processing":[40],"speed.":[41],"These":[42,77],"costs,":[43],"combined":[44],"with":[45],"privacy":[46],"concerns":[47],"around":[48],"cloud-based":[49],"deployments,":[50],"motivated":[52],"research":[53],"into":[54],"running":[55],"LLMs":[56],"on":[57,119],"commodity":[58],"hardware.":[59],"For":[60],"example,":[61],"researchers":[62],"used":[64],"memory":[66],"hierarchy":[67],"to":[68,81,141,146],"boost":[69],"throughput":[70,144],"increasing":[72],"number":[74],"batches.":[76],"studies,":[78],"however,":[79],"tend":[80],"overlook":[82],"or":[83],"inefficiently":[84],"utilize":[85],"additional":[87],"computational":[88],"resources":[89],"provided":[90],"CPU.":[93],"In":[94],"this":[95],"work,":[96],"we":[97],"present":[98],"dynamic":[100],"workload":[101],"allocation":[102],"technique":[103],"that":[104,136],"efficiently":[105],"distributes":[106],"computation":[107],"across":[108],"all":[109],"available":[110],"hardware":[111],"resources.":[112],"The":[113],"proposed":[114],"method":[115],"targets":[116],"decoder-based":[117],"standard":[120],"general-purpose":[121],"hardware,":[122],"effectively":[123],"minimizing":[124],"idle":[125],"periods":[126],"for":[127],"both":[128],"CPU":[130],"GPU.":[133],"Experiments":[134],"show":[135],"our":[137],"approach":[138],"achieves":[139],"up":[140],"30%":[142],"compared":[145],"state":[148],"art,":[151],"regardless":[152],"architecture,":[155],"LLM":[156],"optimizations,":[157],"input":[159],"batch":[160],"sizes.":[161]},"counts_by_year":[],"updated_date":"2026-06-26T06:17:10.115597","created_date":"2026-05-24T00:00:00"}