{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:MD7F23SJLFIYRNU37LZKWBNKET","short_pith_number":"pith:MD7F23SJ","schema_version":"1.0","canonical_sha256":"60fe5d6e49595188b69bfaf2ab05aa24c955a40c39ab8402bbe9b7bd09b62ccd","source":{"kind":"arxiv","id":"2607.16339","version":1},"attestation_state":"computed","paper":{"title":"LaCache: Exact Caching and Precision-Adaptive Inference for Diffusion Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jiqing Zhan, Kun Zhan, Lian Wen, Shuling Yang, Xingru Chen, Yongjia Ma, Zelang Liang","submitted_at":"2026-07-16T10:49:15Z","abstract_excerpt":"Diffusion-based Large Language Models(DLLMs) enable parallel generation via Semi-Autoregressive (SAR) decoding in text generation. However, current methods suffer from severe operator-level redundancy: they recompute the entire sequence during denoising steps, ignoring that the prefix and masked suffix remain invariant within a block. We propose LaCache, a training-free acceleration framework that alleviates this redundancy through lossless caching and mixed precision. Specifically, LaCache employs Lossless State Memoization (LSM) by caching three types of intermediate results: (i) EmbedCache "},"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":"2607.16339","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-16T10:49:15Z","cross_cats_sorted":[],"title_canon_sha256":"0f3e41dfe427a25068ad3e0a6c0a1f71701cf095df9fe127ac8380c97cb8977a","abstract_canon_sha256":"10bb24700870e751cfb61ded1c552c34544e1a6a051ae93d5309fac68bdc45df"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-21T00:20:13.101361Z","signature_b64":"6DxFUvyqNeXZxQ1/Te9Tsb0ZlomfaHEi6WeHOMXg4x66V1bwy4C0Pj7hHqAybWNOXEPXX7occuzogMDunhmUBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"60fe5d6e49595188b69bfaf2ab05aa24c955a40c39ab8402bbe9b7bd09b62ccd","last_reissued_at":"2026-07-21T00:20:13.100446Z","signature_status":"signed_v1","first_computed_at":"2026-07-21T00:20:13.100446Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LaCache: Exact Caching and Precision-Adaptive Inference for Diffusion Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jiqing Zhan, Kun Zhan, Lian Wen, Shuling Yang, Xingru Chen, Yongjia Ma, Zelang Liang","submitted_at":"2026-07-16T10:49:15Z","abstract_excerpt":"Diffusion-based Large Language Models(DLLMs) enable parallel generation via Semi-Autoregressive (SAR) decoding in text generation. However, current methods suffer from severe operator-level redundancy: they recompute the entire sequence during denoising steps, ignoring that the prefix and masked suffix remain invariant within a block. We propose LaCache, a training-free acceleration framework that alleviates this redundancy through lossless caching and mixed precision. Specifically, LaCache employs Lossless State Memoization (LSM) by caching three types of intermediate results: (i) EmbedCache "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.16339","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/2607.16339/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":"2607.16339","created_at":"2026-07-21T00:20:13.100922+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.16339v1","created_at":"2026-07-21T00:20:13.100922+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.16339","created_at":"2026-07-21T00:20:13.100922+00:00"},{"alias_kind":"pith_short_12","alias_value":"MD7F23SJLFIY","created_at":"2026-07-21T00:20:13.100922+00:00"},{"alias_kind":"pith_short_16","alias_value":"MD7F23SJLFIYRNU3","created_at":"2026-07-21T00:20:13.100922+00:00"},{"alias_kind":"pith_short_8","alias_value":"MD7F23SJ","created_at":"2026-07-21T00:20:13.100922+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/MD7F23SJLFIYRNU37LZKWBNKET","json":"https://pith.science/pith/MD7F23SJLFIYRNU37LZKWBNKET.json","graph_json":"https://pith.science/api/pith-number/MD7F23SJLFIYRNU37LZKWBNKET/graph.json","events_json":"https://pith.science/api/pith-number/MD7F23SJLFIYRNU37LZKWBNKET/events.json","paper":"https://pith.science/paper/MD7F23SJ"},"agent_actions":{"view_html":"https://pith.science/pith/MD7F23SJLFIYRNU37LZKWBNKET","download_json":"https://pith.science/pith/MD7F23SJLFIYRNU37LZKWBNKET.json","view_paper":"https://pith.science/paper/MD7F23SJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.16339&json=true","fetch_graph":"https://pith.science/api/pith-number/MD7F23SJLFIYRNU37LZKWBNKET/graph.json","fetch_events":"https://pith.science/api/pith-number/MD7F23SJLFIYRNU37LZKWBNKET/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MD7F23SJLFIYRNU37LZKWBNKET/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MD7F23SJLFIYRNU37LZKWBNKET/action/storage_attestation","attest_author":"https://pith.science/pith/MD7F23SJLFIYRNU37LZKWBNKET/action/author_attestation","sign_citation":"https://pith.science/pith/MD7F23SJLFIYRNU37LZKWBNKET/action/citation_signature","submit_replication":"https://pith.science/pith/MD7F23SJLFIYRNU37LZKWBNKET/action/replication_record"}},"created_at":"2026-07-21T00:20:13.100922+00:00","updated_at":"2026-07-21T00:20:13.100922+00:00"}