{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:IEU6O52PZKZVU2QI7BFHEITW5E","short_pith_number":"pith:IEU6O52P","canonical_record":{"source":{"id":"2410.18038","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-23T17:06:56Z","cross_cats_sorted":["cs.DC"],"title_canon_sha256":"c0140f9e31874333c34ba36536cc611f08cab4ad19540033c2e842154b5bf7f3","abstract_canon_sha256":"df33e4155cdf51791bfbaf91aad18920a61c0c77dedb4264e9569945ec195213"},"schema_version":"1.0"},"canonical_sha256":"4129e7774fcab35a6a08f84a722276e93081aeb196f8308a6e4ee4af92093b1d","source":{"kind":"arxiv","id":"2410.18038","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.18038","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"arxiv_version","alias_value":"2410.18038v2","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.18038","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_12","alias_value":"IEU6O52PZKZV","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_16","alias_value":"IEU6O52PZKZVU2QI","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_8","alias_value":"IEU6O52P","created_at":"2026-07-05T10:14:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:IEU6O52PZKZVU2QI7BFHEITW5E","target":"record","payload":{"canonical_record":{"source":{"id":"2410.18038","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-23T17:06:56Z","cross_cats_sorted":["cs.DC"],"title_canon_sha256":"c0140f9e31874333c34ba36536cc611f08cab4ad19540033c2e842154b5bf7f3","abstract_canon_sha256":"df33e4155cdf51791bfbaf91aad18920a61c0c77dedb4264e9569945ec195213"},"schema_version":"1.0"},"canonical_sha256":"4129e7774fcab35a6a08f84a722276e93081aeb196f8308a6e4ee4af92093b1d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:14:47.039741Z","signature_b64":"XLRc3W2vNwmzSZO/suj2JiRvwV/nEhSTp9xU0arO8OQQGGo5ekKFku2aZqBl2QYI1iEouTXv5ozinMnR4PXHCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4129e7774fcab35a6a08f84a722276e93081aeb196f8308a6e4ee4af92093b1d","last_reissued_at":"2026-07-05T10:14:47.039021Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:14:47.039021Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.18038","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:14:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uNgUNSINrvutO2DCykEKvw6KzLwTusljeC7d2d4v3AixvRgzW9qkHeOZH+vpnYaO+Cs48EmYnl9J3I9CwlaNAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T05:05:37.718237Z"},"content_sha256":"3aad154188ee446ae166fb879110aa77b5267fc9a6e825d27df4f3d774e43316","schema_version":"1.0","event_id":"sha256:3aad154188ee446ae166fb879110aa77b5267fc9a6e825d27df4f3d774e43316"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:IEU6O52PZKZVU2QI7BFHEITW5E","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"POD-Attention: Unlocking Full Prefill-Decode Overlap for Faster LLM Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DC"],"primary_cat":"cs.LG","authors_text":"Aditya K Kamath, Ashish Panwar, Jayashree Mohan, Ramachandran Ramjee, Ramya Prabhu, Simon Peter","submitted_at":"2024-10-23T17:06:56Z","abstract_excerpt":"Each request in LLM inference goes through two phases: compute-bound prefill and memory-bandwidth-bound decode. To improve GPU utilization, recent systems use hybrid batching that combines the prefill and decode phases of different requests into the same batch. This approach optimizes linear operations but remains inefficient for attention computation because existing attention kernels specialize execution independently for the prefill and decode phases.\n  In this paper, we present POD-Attention - the first GPU kernel that efficiently computes attention for hybrid batches. POD-Attention aims t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.18038","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/2410.18038/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:14:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"R/O0n3njUUXKbbtYs0Gc6d2Yw9tLZHd9YHaGHRia8E6y5ugMZr2Tyf/4BWZkA4JJ4BuK+br7KmMzU6GMQ+dtAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T05:05:37.718747Z"},"content_sha256":"791f84d1ad50cf779874745dd27c4d8084eb58796e8d8f87b63386d6d4d50706","schema_version":"1.0","event_id":"sha256:791f84d1ad50cf779874745dd27c4d8084eb58796e8d8f87b63386d6d4d50706"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IEU6O52PZKZVU2QI7BFHEITW5E/bundle.json","state_url":"https://pith.science/pith/IEU6O52PZKZVU2QI7BFHEITW5E/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IEU6O52PZKZVU2QI7BFHEITW5E/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-06T05:05:37Z","links":{"resolver":"https://pith.science/pith/IEU6O52PZKZVU2QI7BFHEITW5E","bundle":"https://pith.science/pith/IEU6O52PZKZVU2QI7BFHEITW5E/bundle.json","state":"https://pith.science/pith/IEU6O52PZKZVU2QI7BFHEITW5E/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IEU6O52PZKZVU2QI7BFHEITW5E/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:IEU6O52PZKZVU2QI7BFHEITW5E","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"df33e4155cdf51791bfbaf91aad18920a61c0c77dedb4264e9569945ec195213","cross_cats_sorted":["cs.DC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-23T17:06:56Z","title_canon_sha256":"c0140f9e31874333c34ba36536cc611f08cab4ad19540033c2e842154b5bf7f3"},"schema_version":"1.0","source":{"id":"2410.18038","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.18038","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"arxiv_version","alias_value":"2410.18038v2","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.18038","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_12","alias_value":"IEU6O52PZKZV","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_16","alias_value":"IEU6O52PZKZVU2QI","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_8","alias_value":"IEU6O52P","created_at":"2026-07-05T10:14:47Z"}],"graph_snapshots":[{"event_id":"sha256:791f84d1ad50cf779874745dd27c4d8084eb58796e8d8f87b63386d6d4d50706","target":"graph","created_at":"2026-07-05T10:14:47Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2410.18038/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Each request in LLM inference goes through two phases: compute-bound prefill and memory-bandwidth-bound decode. To improve GPU utilization, recent systems use hybrid batching that combines the prefill and decode phases of different requests into the same batch. This approach optimizes linear operations but remains inefficient for attention computation because existing attention kernels specialize execution independently for the prefill and decode phases.\n  In this paper, we present POD-Attention - the first GPU kernel that efficiently computes attention for hybrid batches. POD-Attention aims t","authors_text":"Aditya K Kamath, Ashish Panwar, Jayashree Mohan, Ramachandran Ramjee, Ramya Prabhu, Simon Peter","cross_cats":["cs.DC"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-23T17:06:56Z","title":"POD-Attention: Unlocking Full Prefill-Decode Overlap for Faster LLM Inference"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.18038","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:3aad154188ee446ae166fb879110aa77b5267fc9a6e825d27df4f3d774e43316","target":"record","created_at":"2026-07-05T10:14:47Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"df33e4155cdf51791bfbaf91aad18920a61c0c77dedb4264e9569945ec195213","cross_cats_sorted":["cs.DC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-23T17:06:56Z","title_canon_sha256":"c0140f9e31874333c34ba36536cc611f08cab4ad19540033c2e842154b5bf7f3"},"schema_version":"1.0","source":{"id":"2410.18038","kind":"arxiv","version":2}},"canonical_sha256":"4129e7774fcab35a6a08f84a722276e93081aeb196f8308a6e4ee4af92093b1d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4129e7774fcab35a6a08f84a722276e93081aeb196f8308a6e4ee4af92093b1d","first_computed_at":"2026-07-05T10:14:47.039021Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:14:47.039021Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"XLRc3W2vNwmzSZO/suj2JiRvwV/nEhSTp9xU0arO8OQQGGo5ekKFku2aZqBl2QYI1iEouTXv5ozinMnR4PXHCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:14:47.039741Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.18038","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3aad154188ee446ae166fb879110aa77b5267fc9a6e825d27df4f3d774e43316","sha256:791f84d1ad50cf779874745dd27c4d8084eb58796e8d8f87b63386d6d4d50706"],"state_sha256":"272d0353392fdea4a2b818f61165ad286d6b5ca0b78e8faefb4ab31244b915fb"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"B1LpP4rl1fz7u+YElslSxYYZEcJYodSS8A5YkgLhp4CP/hnvvRL8ddrNFAV3swjbQoLXAyCZS6lIEvBYUMFZBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T05:05:37.722665Z","bundle_sha256":"7300f8368154a446db412c7c9988090be2bcad9b3dbdeb53633b574640873fcd"}}