{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:HFW7PH7UXOGHQ3UJDJDGJPABTZ","short_pith_number":"pith:HFW7PH7U","canonical_record":{"source":{"id":"2405.04437","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-07T16:00:32Z","cross_cats_sorted":["cs.OS"],"title_canon_sha256":"be1ae0f2b5fa73be1a9071ab4219f99538b9ffadeccb3e81a542033330af91e4","abstract_canon_sha256":"21b700cc8e17e6b1db74b178c0ef756b4cd3338a50a2391b47e74bbdcead446c"},"schema_version":"1.0"},"canonical_sha256":"396df79ff4bb8c786e891a4664bc019e66736e0c9e4bd7a0c0f6fb3f4cbaca5b","source":{"kind":"arxiv","id":"2405.04437","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.04437","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"arxiv_version","alias_value":"2405.04437v3","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.04437","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"pith_short_12","alias_value":"HFW7PH7UXOGH","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"pith_short_16","alias_value":"HFW7PH7UXOGHQ3UJ","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"pith_short_8","alias_value":"HFW7PH7U","created_at":"2026-07-05T10:06:40Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:HFW7PH7UXOGHQ3UJDJDGJPABTZ","target":"record","payload":{"canonical_record":{"source":{"id":"2405.04437","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-07T16:00:32Z","cross_cats_sorted":["cs.OS"],"title_canon_sha256":"be1ae0f2b5fa73be1a9071ab4219f99538b9ffadeccb3e81a542033330af91e4","abstract_canon_sha256":"21b700cc8e17e6b1db74b178c0ef756b4cd3338a50a2391b47e74bbdcead446c"},"schema_version":"1.0"},"canonical_sha256":"396df79ff4bb8c786e891a4664bc019e66736e0c9e4bd7a0c0f6fb3f4cbaca5b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:06:40.460958Z","signature_b64":"lioAWo/VWGj7iQPnmZen+3utV2AcbQpCskPEn6Ye79fhUek4PDJjVvVJDCem/OCqhuQsX679Y1P6D2MYiUaUAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"396df79ff4bb8c786e891a4664bc019e66736e0c9e4bd7a0c0f6fb3f4cbaca5b","last_reissued_at":"2026-07-05T10:06:40.460465Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:06:40.460465Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.04437","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:06:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Q9900sy+dB8DDNxj0JuSAWnR+ImrAxPj61WihxysiB1TMAkqpZNC6uwadSsvrP+YdajReyDlyBVjZv+KFpWOAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-24T09:22:03.984336Z"},"content_sha256":"115efefeeebd222d81bf892580066281c97cb4445df154b12545a6b48092c9a6","schema_version":"1.0","event_id":"sha256:115efefeeebd222d81bf892580066281c97cb4445df154b12545a6b48092c9a6"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:HFW7PH7UXOGHQ3UJDJDGJPABTZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"vAttention: Dynamic Memory Management for Serving LLMs without PagedAttention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.OS"],"primary_cat":"cs.LG","authors_text":"Ajay Nayak, Ashish Panwar, Jayashree Mohan, Ramachandran Ramjee, Ramya Prabhu","submitted_at":"2024-05-07T16:00:32Z","abstract_excerpt":"PagedAttention is a popular approach for dynamic memory allocation in LLM serving systems. It enables on-demand allocation of GPU memory to mitigate KV cache fragmentation -- a phenomenon that crippled the batch size (and consequently throughput) in prior systems. However, in trying to allocate physical memory at runtime, PagedAttention ends up changing the virtual memory layout of the KV cache from contiguous to non-contiguous. Such a design leads to non-trivial programming and performance overheads.\n  We present vAttention -- an approach that mitigates fragmentation in physical memory while "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.04437","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/2405.04437/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":"2026-07-05T10:06:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iYvO6levLNM6nDT/1HMvrkBJxibf407XOM4Nsa0CIFzBnFGhj7OLcvOV7QjLcHGG08ir7DQyo79DXYaoL07pAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-24T09:22:03.984724Z"},"content_sha256":"4d865e78a9b90c79960e4992bc28de18a5aa33ad9d8596c5cfa5547b153a4eaa","schema_version":"1.0","event_id":"sha256:4d865e78a9b90c79960e4992bc28de18a5aa33ad9d8596c5cfa5547b153a4eaa"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/HFW7PH7UXOGHQ3UJDJDGJPABTZ/bundle.json","state_url":"https://pith.science/pith/HFW7PH7UXOGHQ3UJDJDGJPABTZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/HFW7PH7UXOGHQ3UJDJDGJPABTZ/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":"2026-07-24T09:22:03Z","links":{"resolver":"https://pith.science/pith/HFW7PH7UXOGHQ3UJDJDGJPABTZ","bundle":"https://pith.science/pith/HFW7PH7UXOGHQ3UJDJDGJPABTZ/bundle.json","state":"https://pith.science/pith/HFW7PH7UXOGHQ3UJDJDGJPABTZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/HFW7PH7UXOGHQ3UJDJDGJPABTZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:HFW7PH7UXOGHQ3UJDJDGJPABTZ","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":"21b700cc8e17e6b1db74b178c0ef756b4cd3338a50a2391b47e74bbdcead446c","cross_cats_sorted":["cs.OS"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-07T16:00:32Z","title_canon_sha256":"be1ae0f2b5fa73be1a9071ab4219f99538b9ffadeccb3e81a542033330af91e4"},"schema_version":"1.0","source":{"id":"2405.04437","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.04437","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"arxiv_version","alias_value":"2405.04437v3","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.04437","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"pith_short_12","alias_value":"HFW7PH7UXOGH","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"pith_short_16","alias_value":"HFW7PH7UXOGHQ3UJ","created_at":"2026-07-05T10:06:40Z"},{"alias_kind":"pith_short_8","alias_value":"HFW7PH7U","created_at":"2026-07-05T10:06:40Z"}],"graph_snapshots":[{"event_id":"sha256:4d865e78a9b90c79960e4992bc28de18a5aa33ad9d8596c5cfa5547b153a4eaa","target":"graph","created_at":"2026-07-05T10:06:40Z","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/2405.04437/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"PagedAttention is a popular approach for dynamic memory allocation in LLM serving systems. It enables on-demand allocation of GPU memory to mitigate KV cache fragmentation -- a phenomenon that crippled the batch size (and consequently throughput) in prior systems. However, in trying to allocate physical memory at runtime, PagedAttention ends up changing the virtual memory layout of the KV cache from contiguous to non-contiguous. Such a design leads to non-trivial programming and performance overheads.\n  We present vAttention -- an approach that mitigates fragmentation in physical memory while ","authors_text":"Ajay Nayak, Ashish Panwar, Jayashree Mohan, Ramachandran Ramjee, Ramya Prabhu","cross_cats":["cs.OS"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-07T16:00:32Z","title":"vAttention: Dynamic Memory Management for Serving LLMs without PagedAttention"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.04437","kind":"arxiv","version":3},"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":"sha256:115efefeeebd222d81bf892580066281c97cb4445df154b12545a6b48092c9a6","target":"record","created_at":"2026-07-05T10:06:40Z","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":"21b700cc8e17e6b1db74b178c0ef756b4cd3338a50a2391b47e74bbdcead446c","cross_cats_sorted":["cs.OS"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-07T16:00:32Z","title_canon_sha256":"be1ae0f2b5fa73be1a9071ab4219f99538b9ffadeccb3e81a542033330af91e4"},"schema_version":"1.0","source":{"id":"2405.04437","kind":"arxiv","version":3}},"canonical_sha256":"396df79ff4bb8c786e891a4664bc019e66736e0c9e4bd7a0c0f6fb3f4cbaca5b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"396df79ff4bb8c786e891a4664bc019e66736e0c9e4bd7a0c0f6fb3f4cbaca5b","first_computed_at":"2026-07-05T10:06:40.460465Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:06:40.460465Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"lioAWo/VWGj7iQPnmZen+3utV2AcbQpCskPEn6Ye79fhUek4PDJjVvVJDCem/OCqhuQsX679Y1P6D2MYiUaUAA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:06:40.460958Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.04437","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:115efefeeebd222d81bf892580066281c97cb4445df154b12545a6b48092c9a6","sha256:4d865e78a9b90c79960e4992bc28de18a5aa33ad9d8596c5cfa5547b153a4eaa"],"state_sha256":"1f7f789350c414703528405c59b07cee773d0ec4a1b85cd6875f1d8fd0ddf629"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ziDkbkz9GwSeXg6xBjw4gVtOfx0UXozQXERj7g0DRA6UWGLq6ONtCdqvaOzy84B/CdT4rzkvyiXS2a9U2oGDCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-24T09:22:03.986893Z","bundle_sha256":"f20dfcec520625b09df021b40e1e8d1ba8a2b9624e6225d460a92cbe6a2eae6d"}}