{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IHNTWAOAS5KS4QTD77XQA46AWB","short_pith_number":"pith:IHNTWAOA","schema_version":"1.0","canonical_sha256":"41db3b01c097552e4263ffef0073c0b079073e09af8363af353b8d27446c8c07","source":{"kind":"arxiv","id":"2408.06003","version":3},"attestation_state":"computed","paper":{"title":"LUT Tensor Core: A Software-Hardware Co-Design for LUT-Based Low-Bit LLM Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AR","authors_text":"Fan Yang, Jianyu Wei, Jilong Xue, Lei Wang, Lingxiao Ma, Mao Yang, Naifeng Jing, Shijie Cao, Ting Cao, Zhichen Zeng, Zhiwen Mo","submitted_at":"2024-08-12T08:52:14Z","abstract_excerpt":"Large Language Model (LLM) inference becomes resource-intensive, prompting a shift toward low-bit model weights to reduce the memory footprint and improve efficiency. Such low-bit LLMs necessitate the mixed-precision matrix multiplication (mpGEMM), an important yet underexplored operation involving the multiplication of lower-precision weights with higher-precision activations. Off-the-shelf hardware does not support this operation natively, leading to indirect, thus inefficient, dequantization-based implementations.\n  In this paper, we study the lookup table (LUT)-based approach for mpGEMM an"},"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":"2408.06003","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2024-08-12T08:52:14Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"ec4cf36691154a8de0c8258c099943a91635ac4ab7581fef37df8be5ab08401e","abstract_canon_sha256":"a04b16710bb21ac9db62334872a18b73a702d9898991a27c89c3daac1e83c014"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:43:56.630255Z","signature_b64":"nLmvHUws3wGRdMMztEsziggy0DPqziNDkbftrTi/xjMryXsDVMHMah7c4K84SENbkvYc7g+KCDF2zqDmmbzfDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"41db3b01c097552e4263ffef0073c0b079073e09af8363af353b8d27446c8c07","last_reissued_at":"2026-07-05T11:43:56.629804Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:43:56.629804Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LUT Tensor Core: A Software-Hardware Co-Design for LUT-Based Low-Bit LLM Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AR","authors_text":"Fan Yang, Jianyu Wei, Jilong Xue, Lei Wang, Lingxiao Ma, Mao Yang, Naifeng Jing, Shijie Cao, Ting Cao, Zhichen Zeng, Zhiwen Mo","submitted_at":"2024-08-12T08:52:14Z","abstract_excerpt":"Large Language Model (LLM) inference becomes resource-intensive, prompting a shift toward low-bit model weights to reduce the memory footprint and improve efficiency. Such low-bit LLMs necessitate the mixed-precision matrix multiplication (mpGEMM), an important yet underexplored operation involving the multiplication of lower-precision weights with higher-precision activations. Off-the-shelf hardware does not support this operation natively, leading to indirect, thus inefficient, dequantization-based implementations.\n  In this paper, we study the lookup table (LUT)-based approach for mpGEMM an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.06003","kind":"arxiv","version":3},"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/2408.06003/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":"2408.06003","created_at":"2026-07-05T11:43:56.629859+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.06003v3","created_at":"2026-07-05T11:43:56.629859+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.06003","created_at":"2026-07-05T11:43:56.629859+00:00"},{"alias_kind":"pith_short_12","alias_value":"IHNTWAOAS5KS","created_at":"2026-07-05T11:43:56.629859+00:00"},{"alias_kind":"pith_short_16","alias_value":"IHNTWAOAS5KS4QTD","created_at":"2026-07-05T11:43:56.629859+00:00"},{"alias_kind":"pith_short_8","alias_value":"IHNTWAOA","created_at":"2026-07-05T11:43:56.629859+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.24144","citing_title":"EVA: Accelerating LLM Decoding via an Efficient Vector Quantization Architecture","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2512.06443","citing_title":"Vec-LUT: Vector Table Lookup for Parallel Ultra-Low-Bit LLM Inference on Edge Devices","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IHNTWAOAS5KS4QTD77XQA46AWB","json":"https://pith.science/pith/IHNTWAOAS5KS4QTD77XQA46AWB.json","graph_json":"https://pith.science/api/pith-number/IHNTWAOAS5KS4QTD77XQA46AWB/graph.json","events_json":"https://pith.science/api/pith-number/IHNTWAOAS5KS4QTD77XQA46AWB/events.json","paper":"https://pith.science/paper/IHNTWAOA"},"agent_actions":{"view_html":"https://pith.science/pith/IHNTWAOAS5KS4QTD77XQA46AWB","download_json":"https://pith.science/pith/IHNTWAOAS5KS4QTD77XQA46AWB.json","view_paper":"https://pith.science/paper/IHNTWAOA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.06003&json=true","fetch_graph":"https://pith.science/api/pith-number/IHNTWAOAS5KS4QTD77XQA46AWB/graph.json","fetch_events":"https://pith.science/api/pith-number/IHNTWAOAS5KS4QTD77XQA46AWB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IHNTWAOAS5KS4QTD77XQA46AWB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IHNTWAOAS5KS4QTD77XQA46AWB/action/storage_attestation","attest_author":"https://pith.science/pith/IHNTWAOAS5KS4QTD77XQA46AWB/action/author_attestation","sign_citation":"https://pith.science/pith/IHNTWAOAS5KS4QTD77XQA46AWB/action/citation_signature","submit_replication":"https://pith.science/pith/IHNTWAOAS5KS4QTD77XQA46AWB/action/replication_record"}},"created_at":"2026-07-05T11:43:56.629859+00:00","updated_at":"2026-07-05T11:43:56.629859+00:00"}