{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FZYMMHDANZ2COQWHA5WFXEYHZS","short_pith_number":"pith:FZYMMHDA","schema_version":"1.0","canonical_sha256":"2e70c61c606e742742c7076c5b9307cc961aa08926425b264c51fde5c1a1e54e","source":{"kind":"arxiv","id":"2608.07009","version":1},"attestation_state":"computed","paper":{"title":"HiSparse: Scaling Sparse-Attention Decoding with Hierarchical KV Cache Management","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Christos Kozyrakis, Ruiyang Ma, Tingwei Huang, Zhangheng Huang, Zhiqiang Xie, Ziyi Xu","submitted_at":"2026-08-07T09:22:17Z","abstract_excerpt":"Top-k sparse attention makes long-context LLM decoding cheap to compute: each step reads only a few thousand selected KV entries rather than the full context. Serving systems, however, typically keep the entire KV cache in GPU HBM so that every position stays selectable, so a request's memory bill still grows with its full context length--decoding hits a capacity wall long before it runs out of compute, and a context whose KV cache exceeds HBM cannot be served at all. We present HiSparse, an exact, indexer-agnostic hierarchical KV cache for sparse-attention serving. HiSparse keeps each request"},"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":"2608.07009","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2026-08-07T09:22:17Z","cross_cats_sorted":[],"title_canon_sha256":"2d9a72923933496765185fee95ad1d0f0a86917e9cf87d57be448cb4c87a8414","abstract_canon_sha256":"d81bbf51fc3c6364dc16c803b5ebea9950678b6bcb87cabcdba72b17c98b4c7d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-10T01:12:50.355387Z","signature_b64":"glOc6dNcLSoabmV2ZnKv0kBfugxFobfrUQ/P3WOPASnLT6LBrr8NxIKzF7kpoNGyLAv5IlpGhNQZKqGZ6UFZDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2e70c61c606e742742c7076c5b9307cc961aa08926425b264c51fde5c1a1e54e","last_reissued_at":"2026-08-10T01:12:50.352351Z","signature_status":"signed_v1","first_computed_at":"2026-08-10T01:12:50.352351Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HiSparse: Scaling Sparse-Attention Decoding with Hierarchical KV Cache Management","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Christos Kozyrakis, Ruiyang Ma, Tingwei Huang, Zhangheng Huang, Zhiqiang Xie, Ziyi Xu","submitted_at":"2026-08-07T09:22:17Z","abstract_excerpt":"Top-k sparse attention makes long-context LLM decoding cheap to compute: each step reads only a few thousand selected KV entries rather than the full context. Serving systems, however, typically keep the entire KV cache in GPU HBM so that every position stays selectable, so a request's memory bill still grows with its full context length--decoding hits a capacity wall long before it runs out of compute, and a context whose KV cache exceeds HBM cannot be served at all. We present HiSparse, an exact, indexer-agnostic hierarchical KV cache for sparse-attention serving. HiSparse keeps each request"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.07009","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/2608.07009/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":"2608.07009","created_at":"2026-08-10T01:12:50.353779+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.07009v1","created_at":"2026-08-10T01:12:50.353779+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.07009","created_at":"2026-08-10T01:12:50.353779+00:00"},{"alias_kind":"pith_short_12","alias_value":"FZYMMHDANZ2C","created_at":"2026-08-10T01:12:50.353779+00:00"},{"alias_kind":"pith_short_16","alias_value":"FZYMMHDANZ2COQWH","created_at":"2026-08-10T01:12:50.353779+00:00"},{"alias_kind":"pith_short_8","alias_value":"FZYMMHDA","created_at":"2026-08-10T01:12:50.353779+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/FZYMMHDANZ2COQWHA5WFXEYHZS","json":"https://pith.science/pith/FZYMMHDANZ2COQWHA5WFXEYHZS.json","graph_json":"https://pith.science/api/pith-number/FZYMMHDANZ2COQWHA5WFXEYHZS/graph.json","events_json":"https://pith.science/api/pith-number/FZYMMHDANZ2COQWHA5WFXEYHZS/events.json","paper":"https://pith.science/paper/FZYMMHDA"},"agent_actions":{"view_html":"https://pith.science/pith/FZYMMHDANZ2COQWHA5WFXEYHZS","download_json":"https://pith.science/pith/FZYMMHDANZ2COQWHA5WFXEYHZS.json","view_paper":"https://pith.science/paper/FZYMMHDA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.07009&json=true","fetch_graph":"https://pith.science/api/pith-number/FZYMMHDANZ2COQWHA5WFXEYHZS/graph.json","fetch_events":"https://pith.science/api/pith-number/FZYMMHDANZ2COQWHA5WFXEYHZS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FZYMMHDANZ2COQWHA5WFXEYHZS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FZYMMHDANZ2COQWHA5WFXEYHZS/action/storage_attestation","attest_author":"https://pith.science/pith/FZYMMHDANZ2COQWHA5WFXEYHZS/action/author_attestation","sign_citation":"https://pith.science/pith/FZYMMHDANZ2COQWHA5WFXEYHZS/action/citation_signature","submit_replication":"https://pith.science/pith/FZYMMHDANZ2COQWHA5WFXEYHZS/action/replication_record"}},"created_at":"2026-08-10T01:12:50.353779+00:00","updated_at":"2026-08-10T01:12:50.353779+00:00"}