{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NUKTPVHFO7WTGIELWEVRCMAP5N","short_pith_number":"pith:NUKTPVHF","schema_version":"1.0","canonical_sha256":"6d1537d4e577ed33208bb12b11300feb64843440e4defd6d34c633850a228bf9","source":{"kind":"arxiv","id":"2504.16339","version":1},"attestation_state":"computed","paper":{"title":"Transitive Array: An Efficient GEMM Accelerator with Result Reuse","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Bowen Duan, Chiyue Wei, Cong Guo, Hai Li, Jiaming Tang, Song Han, Yiran Chen","submitted_at":"2025-04-23T01:23:41Z","abstract_excerpt":"Deep Neural Networks (DNNs) and Large Language Models (LLMs) have revolutionized artificial intelligence, yet their deployment faces significant memory and computational challenges, especially in resource-constrained environments. Quantization techniques have mitigated some of these issues by reducing data precision, primarily focusing on General Matrix Multiplication (GEMM). This study introduces a novel sparsity paradigm, transitive sparsity, which leverages the reuse of previously computed results to substantially minimize computational overhead in GEMM operations. By representing transitiv"},"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":"2504.16339","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AR","submitted_at":"2025-04-23T01:23:41Z","cross_cats_sorted":[],"title_canon_sha256":"3d421ec9a7ea85b5789a3ab649261730a13e8b082bb02777bb6eedb122433c02","abstract_canon_sha256":"65d533a6d1a97fd81bfd000d6ff87883da42348b047b4cad91c9d3241fba405c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:52:45.138266Z","signature_b64":"oJuw+/KZQ78U/nI6F8aVPWlfkxB3ECeUWYL8H3IfW0H45vo0W36IgNkSj8sNGjh76HydSgtU4vPjmjP+pDMnCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6d1537d4e577ed33208bb12b11300feb64843440e4defd6d34c633850a228bf9","last_reissued_at":"2026-07-05T10:52:45.137826Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:52:45.137826Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Transitive Array: An Efficient GEMM Accelerator with Result Reuse","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Bowen Duan, Chiyue Wei, Cong Guo, Hai Li, Jiaming Tang, Song Han, Yiran Chen","submitted_at":"2025-04-23T01:23:41Z","abstract_excerpt":"Deep Neural Networks (DNNs) and Large Language Models (LLMs) have revolutionized artificial intelligence, yet their deployment faces significant memory and computational challenges, especially in resource-constrained environments. Quantization techniques have mitigated some of these issues by reducing data precision, primarily focusing on General Matrix Multiplication (GEMM). This study introduces a novel sparsity paradigm, transitive sparsity, which leverages the reuse of previously computed results to substantially minimize computational overhead in GEMM operations. By representing transitiv"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.16339","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/2504.16339/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":"2504.16339","created_at":"2026-07-05T10:52:45.137886+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.16339v1","created_at":"2026-07-05T10:52:45.137886+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.16339","created_at":"2026-07-05T10:52:45.137886+00:00"},{"alias_kind":"pith_short_12","alias_value":"NUKTPVHFO7WT","created_at":"2026-07-05T10:52:45.137886+00:00"},{"alias_kind":"pith_short_16","alias_value":"NUKTPVHFO7WTGIEL","created_at":"2026-07-05T10:52:45.137886+00:00"},{"alias_kind":"pith_short_8","alias_value":"NUKTPVHF","created_at":"2026-07-05T10:52:45.137886+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.09505","citing_title":"Combating the Memory Walls: Optimization Pathways for Long-Context Agentic LLM Inference","ref_index":26,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NUKTPVHFO7WTGIELWEVRCMAP5N","json":"https://pith.science/pith/NUKTPVHFO7WTGIELWEVRCMAP5N.json","graph_json":"https://pith.science/api/pith-number/NUKTPVHFO7WTGIELWEVRCMAP5N/graph.json","events_json":"https://pith.science/api/pith-number/NUKTPVHFO7WTGIELWEVRCMAP5N/events.json","paper":"https://pith.science/paper/NUKTPVHF"},"agent_actions":{"view_html":"https://pith.science/pith/NUKTPVHFO7WTGIELWEVRCMAP5N","download_json":"https://pith.science/pith/NUKTPVHFO7WTGIELWEVRCMAP5N.json","view_paper":"https://pith.science/paper/NUKTPVHF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.16339&json=true","fetch_graph":"https://pith.science/api/pith-number/NUKTPVHFO7WTGIELWEVRCMAP5N/graph.json","fetch_events":"https://pith.science/api/pith-number/NUKTPVHFO7WTGIELWEVRCMAP5N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NUKTPVHFO7WTGIELWEVRCMAP5N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NUKTPVHFO7WTGIELWEVRCMAP5N/action/storage_attestation","attest_author":"https://pith.science/pith/NUKTPVHFO7WTGIELWEVRCMAP5N/action/author_attestation","sign_citation":"https://pith.science/pith/NUKTPVHFO7WTGIELWEVRCMAP5N/action/citation_signature","submit_replication":"https://pith.science/pith/NUKTPVHFO7WTGIELWEVRCMAP5N/action/replication_record"}},"created_at":"2026-07-05T10:52:45.137886+00:00","updated_at":"2026-07-05T10:52:45.137886+00:00"}