{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:DBMU22BD6Y5B6KLLNCLO5GM4RY","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":"aef4dc47dbab3e2d5063665631129f3e190d2bc4cba239b5a21522504ab47a58","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2024-09-08T06:06:44Z","title_canon_sha256":"50664bfba5c194046152cf462569a6f6528f5d820074af7660123f8a8d568c2b"},"schema_version":"1.0","source":{"id":"2409.04992","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2409.04992","created_at":"2026-07-05T09:04:42Z"},{"alias_kind":"arxiv_version","alias_value":"2409.04992v1","created_at":"2026-07-05T09:04:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.04992","created_at":"2026-07-05T09:04:42Z"},{"alias_kind":"pith_short_12","alias_value":"DBMU22BD6Y5B","created_at":"2026-07-05T09:04:42Z"},{"alias_kind":"pith_short_16","alias_value":"DBMU22BD6Y5B6KLL","created_at":"2026-07-05T09:04:42Z"},{"alias_kind":"pith_short_8","alias_value":"DBMU22BD","created_at":"2026-07-05T09:04:42Z"}],"graph_snapshots":[{"event_id":"sha256:8c0f8d74db4096c2bbdf1457a1bc2e788336f9b6416204a1de92d9b97ab4ec9e","target":"graph","created_at":"2026-07-05T09:04:42Z","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/2409.04992/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The widespread of Large Language Models (LLMs) marks a significant milestone in generative AI. Nevertheless, the increasing context length and batch size in offline LLM inference escalate the memory requirement of the key-value (KV) cache, which imposes a huge burden on the GPU VRAM, especially for resource-constraint scenarios (e.g., edge computing and personal devices). Several cost-effective solutions leverage host memory or SSDs to reduce storage costs for offline inference scenarios and improve the throughput. Nevertheless, they suffer from significant performance penalties imposed by int","authors_text":"Endian Li, Jie Zhang, Ke Zhou, Qiao Li, Shengwen Liang, Xiaolin Wang, Xiurui Pan, Yingwei Luo, Yizhou Shan","cross_cats":["cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2024-09-08T06:06:44Z","title":"InstInfer: In-Storage Attention Offloading for Cost-Effective Long-Context LLM Inference"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.04992","kind":"arxiv","version":1},"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":"sha256:5b575360bfd192b71d48b6423ec7a86505003f9c1aee6704a4edc18d80b75b9b","target":"record","created_at":"2026-07-05T09:04:42Z","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":"aef4dc47dbab3e2d5063665631129f3e190d2bc4cba239b5a21522504ab47a58","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2024-09-08T06:06:44Z","title_canon_sha256":"50664bfba5c194046152cf462569a6f6528f5d820074af7660123f8a8d568c2b"},"schema_version":"1.0","source":{"id":"2409.04992","kind":"arxiv","version":1}},"canonical_sha256":"18594d6823f63a1f296b6896ee999c8e3ac7147e465d63e503d3af6668d80e24","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"18594d6823f63a1f296b6896ee999c8e3ac7147e465d63e503d3af6668d80e24","first_computed_at":"2026-07-05T09:04:42.659810Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:04:42.659810Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"zpkE4jqLx0S8dkumUC0xN666w5wOypibOtgs8UJKykGtBBn4ejVHZTa/3EecJpCweBmXJq7S6sCYlmk7ItmtCg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:04:42.660282Z","signed_message":"canonical_sha256_bytes"},"source_id":"2409.04992","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5b575360bfd192b71d48b6423ec7a86505003f9c1aee6704a4edc18d80b75b9b","sha256:8c0f8d74db4096c2bbdf1457a1bc2e788336f9b6416204a1de92d9b97ab4ec9e"],"state_sha256":"21a0e3f71d0dddb41f29bb2f864f4195a326c0959854f008f2fe90b56c5ae64a"}