{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:GWEQCPATMS57HL7WKADNHO265D","short_pith_number":"pith:GWEQCPAT","schema_version":"1.0","canonical_sha256":"3589013c1364bbf3aff65006d3bb5ee8d5b96c432c2c8f3cbb1fa55806779462","source":{"kind":"arxiv","id":"2402.10076","version":1},"attestation_state":"computed","paper":{"title":"QUICK: Quantization-aware Interleaving and Conflict-free Kernel for efficient LLM inference","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Daehyun Ahn, Hyungjun Kim, Jiwoong Choi, Jongho Lee, Minkyu Kim, Sarang Kim, Taesu Kim","submitted_at":"2024-02-15T16:38:41Z","abstract_excerpt":"We introduce QUICK, a group of novel optimized CUDA kernels for the efficient inference of quantized Large Language Models (LLMs). QUICK addresses the shared memory bank-conflict problem of state-of-the-art mixed precision matrix multiplication kernels. Our method interleaves the quantized weight matrices of LLMs offline to skip the shared memory write-back after the dequantization. We demonstrate up to 1.91x speedup over existing kernels of AutoAWQ on larger batches and up to 1.94x throughput gain on representative LLM models on various NVIDIA GPU devices."},"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":"2402.10076","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-15T16:38:41Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"7dcf9d1cb9ad23cb51910da4a39ec8b4d24031c9870d99eb05a9d209254c9d9c","abstract_canon_sha256":"335a287db6d1d2054b5d55d746ce7e18c4de651bc9d4694ec145e27dbc8238f2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:45:41.036665Z","signature_b64":"ADZTT7vV6uxc8vrOdQlisZvYKIHD091DDKGPQM2yQGuYDmVlyQOyNMHGDkfmO20viyQ7Y95jTA7nKNhNjlvuBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3589013c1364bbf3aff65006d3bb5ee8d5b96c432c2c8f3cbb1fa55806779462","last_reissued_at":"2026-07-05T07:45:41.036220Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:45:41.036220Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"QUICK: Quantization-aware Interleaving and Conflict-free Kernel for efficient LLM inference","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Daehyun Ahn, Hyungjun Kim, Jiwoong Choi, Jongho Lee, Minkyu Kim, Sarang Kim, Taesu Kim","submitted_at":"2024-02-15T16:38:41Z","abstract_excerpt":"We introduce QUICK, a group of novel optimized CUDA kernels for the efficient inference of quantized Large Language Models (LLMs). QUICK addresses the shared memory bank-conflict problem of state-of-the-art mixed precision matrix multiplication kernels. Our method interleaves the quantized weight matrices of LLMs offline to skip the shared memory write-back after the dequantization. We demonstrate up to 1.91x speedup over existing kernels of AutoAWQ on larger batches and up to 1.94x throughput gain on representative LLM models on various NVIDIA GPU devices."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.10076","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/2402.10076/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":"2402.10076","created_at":"2026-07-05T07:45:41.036279+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.10076v1","created_at":"2026-07-05T07:45:41.036279+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.10076","created_at":"2026-07-05T07:45:41.036279+00:00"},{"alias_kind":"pith_short_12","alias_value":"GWEQCPATMS57","created_at":"2026-07-05T07:45:41.036279+00:00"},{"alias_kind":"pith_short_16","alias_value":"GWEQCPATMS57HL7W","created_at":"2026-07-05T07:45:41.036279+00:00"},{"alias_kind":"pith_short_8","alias_value":"GWEQCPAT","created_at":"2026-07-05T07:45:41.036279+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2508.15601","citing_title":"LMDeploy Accelerates Mixed-Precision LLM Inference with TurboMind","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14789","citing_title":"A Comparative Study of CNN Optimization Methods for Edge AI: Exploring the Role of Early Exits","ref_index":65,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GWEQCPATMS57HL7WKADNHO265D","json":"https://pith.science/pith/GWEQCPATMS57HL7WKADNHO265D.json","graph_json":"https://pith.science/api/pith-number/GWEQCPATMS57HL7WKADNHO265D/graph.json","events_json":"https://pith.science/api/pith-number/GWEQCPATMS57HL7WKADNHO265D/events.json","paper":"https://pith.science/paper/GWEQCPAT"},"agent_actions":{"view_html":"https://pith.science/pith/GWEQCPATMS57HL7WKADNHO265D","download_json":"https://pith.science/pith/GWEQCPATMS57HL7WKADNHO265D.json","view_paper":"https://pith.science/paper/GWEQCPAT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.10076&json=true","fetch_graph":"https://pith.science/api/pith-number/GWEQCPATMS57HL7WKADNHO265D/graph.json","fetch_events":"https://pith.science/api/pith-number/GWEQCPATMS57HL7WKADNHO265D/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GWEQCPATMS57HL7WKADNHO265D/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GWEQCPATMS57HL7WKADNHO265D/action/storage_attestation","attest_author":"https://pith.science/pith/GWEQCPATMS57HL7WKADNHO265D/action/author_attestation","sign_citation":"https://pith.science/pith/GWEQCPATMS57HL7WKADNHO265D/action/citation_signature","submit_replication":"https://pith.science/pith/GWEQCPATMS57HL7WKADNHO265D/action/replication_record"}},"created_at":"2026-07-05T07:45:41.036279+00:00","updated_at":"2026-07-05T07:45:41.036279+00:00"}