{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:PGIGHK2PDSKF3DQ2NDFAGOLWXU","short_pith_number":"pith:PGIGHK2P","schema_version":"1.0","canonical_sha256":"799063ab4f1c945d8e1a68ca033976bd07351d1648b275cf389afcd85ba6b764","source":{"kind":"arxiv","id":"2507.12205","version":1},"attestation_state":"computed","paper":{"title":"Toward Efficient SpMV in Sparse LLMs via Block Extraction and Compressed Storage","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Guangzhong Sun, Jingwei Sun, Junqing Lin, Mingge Lu","submitted_at":"2025-07-16T13:04:06Z","abstract_excerpt":"Sparse Matrix-Vector Multiplication (SpMV) has become a critical performance bottleneck in the local deployment of sparse Large Language Models (LLMs), where inference predominantly operates on workloads during the decoder phase with a batch size of one. Existing SpMV kernels and sparse matrix formats, originally designed for scientific computing, fail to exploit the unique structure patterns inherent in sparse LLMs, resulting in suboptimal performance and excessive storage overhead. This paper presents EC-SpMV, a GPU-optimized SpMV approach for accelerating sparse LLM inference. EC-SpMV intro"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2507.12205","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-07-16T13:04:06Z","cross_cats_sorted":[],"title_canon_sha256":"cf2bc68e0d9a18c18f9cb318fdf3709eba1fa6b5e36f6a6a4e536ea08fb062b6","abstract_canon_sha256":"db903447a22435cd73e3be3af48d6025eb04a6f8d9607f7f7deac965ec4bf331"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:38:16.565028Z","signature_b64":"BnMivVHyNQHJ3HV0jkxl51nAByy8VnyrBu/zLIdBN3QCp5ioaE4FguwvErA/f1VbyEUSXIaoWtjqqDdrtLpiDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"799063ab4f1c945d8e1a68ca033976bd07351d1648b275cf389afcd85ba6b764","last_reissued_at":"2026-07-05T11:38:16.564537Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:38:16.564537Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Toward Efficient SpMV in Sparse LLMs via Block Extraction and Compressed Storage","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Guangzhong Sun, Jingwei Sun, Junqing Lin, Mingge Lu","submitted_at":"2025-07-16T13:04:06Z","abstract_excerpt":"Sparse Matrix-Vector Multiplication (SpMV) has become a critical performance bottleneck in the local deployment of sparse Large Language Models (LLMs), where inference predominantly operates on workloads during the decoder phase with a batch size of one. Existing SpMV kernels and sparse matrix formats, originally designed for scientific computing, fail to exploit the unique structure patterns inherent in sparse LLMs, resulting in suboptimal performance and excessive storage overhead. This paper presents EC-SpMV, a GPU-optimized SpMV approach for accelerating sparse LLM inference. EC-SpMV intro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.12205","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2507.12205/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2507.12205","created_at":"2026-07-05T11:38:16.564595+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.12205v1","created_at":"2026-07-05T11:38:16.564595+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.12205","created_at":"2026-07-05T11:38:16.564595+00:00"},{"alias_kind":"pith_short_12","alias_value":"PGIGHK2PDSKF","created_at":"2026-07-05T11:38:16.564595+00:00"},{"alias_kind":"pith_short_16","alias_value":"PGIGHK2PDSKF3DQ2","created_at":"2026-07-05T11:38:16.564595+00:00"},{"alias_kind":"pith_short_8","alias_value":"PGIGHK2P","created_at":"2026-07-05T11:38:16.564595+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PGIGHK2PDSKF3DQ2NDFAGOLWXU","json":"https://pith.science/pith/PGIGHK2PDSKF3DQ2NDFAGOLWXU.json","graph_json":"https://pith.science/api/pith-number/PGIGHK2PDSKF3DQ2NDFAGOLWXU/graph.json","events_json":"https://pith.science/api/pith-number/PGIGHK2PDSKF3DQ2NDFAGOLWXU/events.json","paper":"https://pith.science/paper/PGIGHK2P"},"agent_actions":{"view_html":"https://pith.science/pith/PGIGHK2PDSKF3DQ2NDFAGOLWXU","download_json":"https://pith.science/pith/PGIGHK2PDSKF3DQ2NDFAGOLWXU.json","view_paper":"https://pith.science/paper/PGIGHK2P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.12205&json=true","fetch_graph":"https://pith.science/api/pith-number/PGIGHK2PDSKF3DQ2NDFAGOLWXU/graph.json","fetch_events":"https://pith.science/api/pith-number/PGIGHK2PDSKF3DQ2NDFAGOLWXU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PGIGHK2PDSKF3DQ2NDFAGOLWXU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PGIGHK2PDSKF3DQ2NDFAGOLWXU/action/storage_attestation","attest_author":"https://pith.science/pith/PGIGHK2PDSKF3DQ2NDFAGOLWXU/action/author_attestation","sign_citation":"https://pith.science/pith/PGIGHK2PDSKF3DQ2NDFAGOLWXU/action/citation_signature","submit_replication":"https://pith.science/pith/PGIGHK2PDSKF3DQ2NDFAGOLWXU/action/replication_record"}},"created_at":"2026-07-05T11:38:16.564595+00:00","updated_at":"2026-07-05T11:38:16.564595+00:00"}