{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BLHNG7SNPYUHZA6DLFAYSSKMLK","short_pith_number":"pith:BLHNG7SN","schema_version":"1.0","canonical_sha256":"0aced37e4d7e287c83c3594189494c5a884d3f5626452e77c85e2312000f7792","source":{"kind":"arxiv","id":"2409.17870","version":2},"attestation_state":"computed","paper":{"title":"Efficient Arbitrary Precision Acceleration for Large Language Models on GPU Tensor Cores","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.AR"],"primary_cat":"cs.LG","authors_text":"Chao Fang, Haikuo Shao, Shaobo Ma, Zhongfeng Wang","submitted_at":"2024-09-26T14:17:58Z","abstract_excerpt":"Large language models (LLMs) have been widely applied but face challenges in efficient inference. While quantization methods reduce computational demands, ultra-low bit quantization with arbitrary precision is hindered by limited GPU Tensor Core support and inefficient memory management, leading to suboptimal acceleration. To address these challenges, we propose a comprehensive acceleration scheme for arbitrary precision LLMs. At its core, we introduce a novel bipolar-INT data format that facilitates parallel computing and supports symmetric quantization, effectively reducing data redundancy. "},"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":"2409.17870","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-09-26T14:17:58Z","cross_cats_sorted":["cs.AI","cs.AR"],"title_canon_sha256":"790c599e253aebd28d3257d1afc486b7b922af7b24af61e7980ead275f979bef","abstract_canon_sha256":"4a0a54ee058cf74beedfc6829a7309048fe6c09315d9c1b0f773bd01287c7f05"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:30:23.517465Z","signature_b64":"7kSCN39E11eo88Q6mPkGWO2L6KhXFcc49VuCAO8NOLZyeFIucZjajUBk4WYbu8swYybySZZftE+FKM1w/+BsAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0aced37e4d7e287c83c3594189494c5a884d3f5626452e77c85e2312000f7792","last_reissued_at":"2026-07-05T10:30:23.516724Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:30:23.516724Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient Arbitrary Precision Acceleration for Large Language Models on GPU Tensor Cores","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.AR"],"primary_cat":"cs.LG","authors_text":"Chao Fang, Haikuo Shao, Shaobo Ma, Zhongfeng Wang","submitted_at":"2024-09-26T14:17:58Z","abstract_excerpt":"Large language models (LLMs) have been widely applied but face challenges in efficient inference. While quantization methods reduce computational demands, ultra-low bit quantization with arbitrary precision is hindered by limited GPU Tensor Core support and inefficient memory management, leading to suboptimal acceleration. To address these challenges, we propose a comprehensive acceleration scheme for arbitrary precision LLMs. At its core, we introduce a novel bipolar-INT data format that facilitates parallel computing and supports symmetric quantization, effectively reducing data redundancy. "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.17870","kind":"arxiv","version":2},"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/2409.17870/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":"2409.17870","created_at":"2026-07-05T10:30:23.516798+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.17870v2","created_at":"2026-07-05T10:30:23.516798+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.17870","created_at":"2026-07-05T10:30:23.516798+00:00"},{"alias_kind":"pith_short_12","alias_value":"BLHNG7SNPYUH","created_at":"2026-07-05T10:30:23.516798+00:00"},{"alias_kind":"pith_short_16","alias_value":"BLHNG7SNPYUHZA6D","created_at":"2026-07-05T10:30:23.516798+00:00"},{"alias_kind":"pith_short_8","alias_value":"BLHNG7SN","created_at":"2026-07-05T10:30:23.516798+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.15982","citing_title":"Anda: Unlocking Efficient LLM Inference with a Variable-Length Grouped Activation Data Format","ref_index":57,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BLHNG7SNPYUHZA6DLFAYSSKMLK","json":"https://pith.science/pith/BLHNG7SNPYUHZA6DLFAYSSKMLK.json","graph_json":"https://pith.science/api/pith-number/BLHNG7SNPYUHZA6DLFAYSSKMLK/graph.json","events_json":"https://pith.science/api/pith-number/BLHNG7SNPYUHZA6DLFAYSSKMLK/events.json","paper":"https://pith.science/paper/BLHNG7SN"},"agent_actions":{"view_html":"https://pith.science/pith/BLHNG7SNPYUHZA6DLFAYSSKMLK","download_json":"https://pith.science/pith/BLHNG7SNPYUHZA6DLFAYSSKMLK.json","view_paper":"https://pith.science/paper/BLHNG7SN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.17870&json=true","fetch_graph":"https://pith.science/api/pith-number/BLHNG7SNPYUHZA6DLFAYSSKMLK/graph.json","fetch_events":"https://pith.science/api/pith-number/BLHNG7SNPYUHZA6DLFAYSSKMLK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BLHNG7SNPYUHZA6DLFAYSSKMLK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BLHNG7SNPYUHZA6DLFAYSSKMLK/action/storage_attestation","attest_author":"https://pith.science/pith/BLHNG7SNPYUHZA6DLFAYSSKMLK/action/author_attestation","sign_citation":"https://pith.science/pith/BLHNG7SNPYUHZA6DLFAYSSKMLK/action/citation_signature","submit_replication":"https://pith.science/pith/BLHNG7SNPYUHZA6DLFAYSSKMLK/action/replication_record"}},"created_at":"2026-07-05T10:30:23.516798+00:00","updated_at":"2026-07-05T10:30:23.516798+00:00"}