{"id":"https://openalex.org/W4415330726","doi":"https://doi.org/10.1109/asp-dac66049.2026.11420607","title":"SnipSnap: A Joint Compression Format and Dataflow Co-Optimization Framework for Efficient Sparse LLM Accelerator Design","display_name":"SnipSnap: A Joint Compression Format and Dataflow Co-Optimization Framework for Efficient Sparse LLM Accelerator Design","publication_year":2026,"publication_date":"2026-01-19","ids":{"openalex":"https://openalex.org/W4415330726","doi":"https://doi.org/10.1109/asp-dac66049.2026.11420607"},"language":"en","primary_location":{"id":"doi:10.1109/asp-dac66049.2026.11420607","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asp-dac66049.2026.11420607","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 31st Asia and South Pacific Design Automation Conference (ASP-DAC)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2509.17072","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5111126312","display_name":"Junyi Wu","orcid":"https://orcid.org/0000-0002-7417-3356"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Junyi Wu","raw_affiliation_strings":["Nanjing University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanjing University","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103213382","display_name":"Chao Fang","orcid":"https://orcid.org/0000-0003-4967-0411"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chao Fang","raw_affiliation_strings":["Nanjing University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanjing University","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100697000","display_name":"Zhongfeng Wang","orcid":"https://orcid.org/0000-0003-0402-7334"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhongfeng Wang","raw_affiliation_strings":["Nanjing University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanjing University","institution_ids":["https://openalex.org/I881766915"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I881766915"],"apc_list":null,"apc_paid":null,"fwci":10.2798,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.92991071,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"865","last_page":"871"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.984000027179718,"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":0.984000027179718,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.9833999872207642,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T10901","display_name":"Advanced Data Compression Techniques","score":0.9775999784469604,"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/dataflow","display_name":"Dataflow","score":0.873199999332428},{"id":"https://openalex.org/keywords/workflow","display_name":"Workflow","score":0.6689000129699707},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.6650000214576721},{"id":"https://openalex.org/keywords/joint","display_name":"Joint (building)","score":0.6107000112533569},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.5687000155448914},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.5608000159263611},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5268999934196472},{"id":"https://openalex.org/keywords/encoding","display_name":"Encoding (memory)","score":0.5257999897003174},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.45500001311302185}],"concepts":[{"id":"https://openalex.org/C96324660","wikidata":"https://www.wikidata.org/wiki/Q205446","display_name":"Dataflow","level":2,"score":0.873199999332428},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8327999711036682},{"id":"https://openalex.org/C177212765","wikidata":"https://www.wikidata.org/wiki/Q627335","display_name":"Workflow","level":2,"score":0.6689000129699707},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.6650000214576721},{"id":"https://openalex.org/C18555067","wikidata":"https://www.wikidata.org/wiki/Q8375051","display_name":"Joint (building)","level":2,"score":0.6107000112533569},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.5687000155448914},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.5608000159263611},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5268999934196472},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.5257999897003174},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.45500001311302185},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.4088999927043915},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.37630000710487366},{"id":"https://openalex.org/C25797200","wikidata":"https://www.wikidata.org/wiki/Q828137","display_name":"Compression ratio","level":3,"score":0.37529999017715454},{"id":"https://openalex.org/C2781039887","wikidata":"https://www.wikidata.org/wiki/Q1391724","display_name":"Factor (programming language)","level":2,"score":0.3643999993801117},{"id":"https://openalex.org/C169590947","wikidata":"https://www.wikidata.org/wiki/Q47506","display_name":"Compiler","level":2,"score":0.361299991607666},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.3612000048160553},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.36070001125335693},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3596999943256378},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3594000041484833},{"id":"https://openalex.org/C489000","wikidata":"https://www.wikidata.org/wiki/Q747385","display_name":"Data flow diagram","level":2,"score":0.3564000129699707},{"id":"https://openalex.org/C168781493","wikidata":"https://www.wikidata.org/wiki/Q80585","display_name":"Associative array","level":2,"score":0.34929999709129333},{"id":"https://openalex.org/C186370098","wikidata":"https://www.wikidata.org/wiki/Q442787","display_name":"Energy (signal processing)","level":2,"score":0.3395000100135803},{"id":"https://openalex.org/C97250363","wikidata":"https://www.wikidata.org/wiki/Q235557","display_name":"File format","level":2,"score":0.3393999934196472},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.33149999380111694},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.2939999997615814},{"id":"https://openalex.org/C162319229","wikidata":"https://www.wikidata.org/wiki/Q175263","display_name":"Data structure","level":2,"score":0.2939000129699707},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.2849000096321106},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.26019999384880066},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.25949999690055847},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.25929999351501465},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.2563000023365021},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.2549999952316284}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/asp-dac66049.2026.11420607","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asp-dac66049.2026.11420607","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 31st Asia and South Pacific Design Automation Conference (ASP-DAC)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2509.17072","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2509.17072","pdf_url":"https://arxiv.org/pdf/2509.17072","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2509.17072","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2509.17072","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":"pmh:oai:arXiv.org:2509.17072","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2509.17072","pdf_url":"https://arxiv.org/pdf/2509.17072","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4415330726.pdf","grobid_xml":"https://content.openalex.org/works/W4415330726.grobid-xml"},"referenced_works_count":35,"referenced_works":["https://openalex.org/W4415797802","https://openalex.org/W2289252105","https://openalex.org/W2945146780","https://openalex.org/W4291653336","https://openalex.org/W4403211793","https://openalex.org/W2979439447","https://openalex.org/W4312569226","https://openalex.org/W4393406875","https://openalex.org/W4360606782","https://openalex.org/W4285130446","https://openalex.org/W4386902741","https://openalex.org/W2904902077","https://openalex.org/W3105802176","https://openalex.org/W4408183166","https://openalex.org/W2940862705","https://openalex.org/W3132942233","https://openalex.org/W4401211642","https://openalex.org/W4285335127","https://openalex.org/W4401568161","https://openalex.org/W4308083753","https://openalex.org/W4411446407","https://openalex.org/W2997929983","https://openalex.org/W2979644612","https://openalex.org/W4404057367","https://openalex.org/W4214686755","https://openalex.org/W2980186997","https://openalex.org/W4240168186","https://openalex.org/W2625457103","https://openalex.org/W4327930477","https://openalex.org/W4392248665","https://openalex.org/W4416310693","https://openalex.org/W2194775991","https://openalex.org/W3016542674","https://openalex.org/W3185702163","https://openalex.org/W3114479342"],"related_works":[],"abstract_inverted_index":{"The":[0],"growing":[1],"scale":[2],"of":[3],"large":[4],"language":[5],"models":[6],"(LLMs)":[7],"has":[8],"intensified":[9],"demands":[10],"on":[11,43],"computation":[12],"and":[13,53,87,97,113,118],"memory,":[14],"making":[15],"efficient":[16,58],"inference":[17],"a":[18,37,49,66,89],"key":[19,38],"challenge.":[20],"While":[21],"sparsity":[22,42],"can":[23],"reduce":[24],"these":[25],"costs,":[26],"existing":[27],"design":[28,74],"space":[29],"exploration":[30],"(DSE)":[31],"frameworks":[32],"often":[33],"overlook":[34],"compression":[35,51,68,79,98],"formats,":[36],"factor":[39],"for":[40,57,81],"leveraging":[41],"accelerators.":[44],"This":[45],"paper":[46],"proposes":[47],"SnipSnap,":[48],"joint":[50],"format":[52,69,108],"dataflow":[54,96],"co-optimization":[55],"framework":[56],"sparse":[59],"LLM":[60],"accelerator":[61],"design.":[62],"SnipSnap":[63,100],"introduces:":[64],"(1)":[65],"hierarchical":[67],"encoding":[70],"to":[71],"expand":[72],"the":[73],"space;":[75],"(2)":[76],"an":[77],"adaptive":[78],"engine":[80],"selecting":[82],"formats":[83],"under":[84],"diverse":[85],"sparsity;":[86],"(3)":[88],"progressive":[90],"co-search":[91],"workflow":[92],"that":[93],"jointly":[94],"optimizes":[95],"formats.":[99],"achieves":[101],"18.24%":[102],"average":[103],"memory":[104],"energy":[105],"savings":[106],"via":[107],"optimization,":[109],"along":[110],"with":[111],"2248.3$\\times$":[112],"21.0$\\times$":[114],"speedups":[115],"over":[116],"Sparseloop":[117],"DiMO-Sparse":[119],"frameworks,":[120],"respectively.":[121]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-08-18T07:49:30.821534","created_date":"2025-10-19T00:00:00"}