{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:RXAIV4XQ7K6DKJYTDVFDUU7W7M","short_pith_number":"pith:RXAIV4XQ","schema_version":"1.0","canonical_sha256":"8dc08af2f0fabc3527131d4a3a53f6fb090167f30f3b05cb62de36f30624765b","source":{"kind":"arxiv","id":"2606.01502","version":1},"attestation_state":"computed","paper":{"title":"Move the Query, Not the Cache: Characterizing Cross-Instance Latent Attention Redistribution Across GPU Fabrics","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.NI"],"primary_cat":"cs.DC","authors_text":"Bole Ma, Gerhard Wellein, Harald K\\\"ostler, Jan Eitzinger","submitted_at":"2026-05-31T23:53:24Z","abstract_excerpt":"Frontier LLMs increasingly decide what a query attends to with a sparse-attention indexer that picks a few KV-cache blocks per query: attention's unit is now a small, reusable chunk. Agentic workloads hammer it: many sub-agents query one large codebase, reusing the same blocks. When that corpus outgrows one GPU it is partitioned across instances, so a query and the blocks it selects often sit on different GPUs: answering it means attention across instances. The reflex of prior cross-instance KV systems is to move the cache: pull the selected blocks to the requester. Multi-head Latent Attention"},"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":"2606.01502","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2026-05-31T23:53:24Z","cross_cats_sorted":["cs.AI","cs.NI"],"title_canon_sha256":"0ceff7097be37016cf7b71518f9688e626b8ee2c58151aa5f9248a0089374e5f","abstract_canon_sha256":"e80ecd9d93e55fb4d5018bb09becdcbc8fbc155eadc73c6b4b9990a8d17d909d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:34.924122Z","signature_b64":"5Jgo4kSBa2zdhk6M4Yh3p9PbIRo8vJQjcsP9daqV00vSv6R3/71XO/j7dCIvnqwg3n1wqbQbeCcPi88MymWFAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8dc08af2f0fabc3527131d4a3a53f6fb090167f30f3b05cb62de36f30624765b","last_reissued_at":"2026-06-02T02:04:34.923720Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:34.923720Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Move the Query, Not the Cache: Characterizing Cross-Instance Latent Attention Redistribution Across GPU Fabrics","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.NI"],"primary_cat":"cs.DC","authors_text":"Bole Ma, Gerhard Wellein, Harald K\\\"ostler, Jan Eitzinger","submitted_at":"2026-05-31T23:53:24Z","abstract_excerpt":"Frontier LLMs increasingly decide what a query attends to with a sparse-attention indexer that picks a few KV-cache blocks per query: attention's unit is now a small, reusable chunk. Agentic workloads hammer it: many sub-agents query one large codebase, reusing the same blocks. When that corpus outgrows one GPU it is partitioned across instances, so a query and the blocks it selects often sit on different GPUs: answering it means attention across instances. The reflex of prior cross-instance KV systems is to move the cache: pull the selected blocks to the requester. Multi-head Latent Attention"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.01502","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/2606.01502/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":"2606.01502","created_at":"2026-06-02T02:04:34.923779+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.01502v1","created_at":"2026-06-02T02:04:34.923779+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.01502","created_at":"2026-06-02T02:04:34.923779+00:00"},{"alias_kind":"pith_short_12","alias_value":"RXAIV4XQ7K6D","created_at":"2026-06-02T02:04:34.923779+00:00"},{"alias_kind":"pith_short_16","alias_value":"RXAIV4XQ7K6DKJYT","created_at":"2026-06-02T02:04:34.923779+00:00"},{"alias_kind":"pith_short_8","alias_value":"RXAIV4XQ","created_at":"2026-06-02T02:04:34.923779+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/RXAIV4XQ7K6DKJYTDVFDUU7W7M","json":"https://pith.science/pith/RXAIV4XQ7K6DKJYTDVFDUU7W7M.json","graph_json":"https://pith.science/api/pith-number/RXAIV4XQ7K6DKJYTDVFDUU7W7M/graph.json","events_json":"https://pith.science/api/pith-number/RXAIV4XQ7K6DKJYTDVFDUU7W7M/events.json","paper":"https://pith.science/paper/RXAIV4XQ"},"agent_actions":{"view_html":"https://pith.science/pith/RXAIV4XQ7K6DKJYTDVFDUU7W7M","download_json":"https://pith.science/pith/RXAIV4XQ7K6DKJYTDVFDUU7W7M.json","view_paper":"https://pith.science/paper/RXAIV4XQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.01502&json=true","fetch_graph":"https://pith.science/api/pith-number/RXAIV4XQ7K6DKJYTDVFDUU7W7M/graph.json","fetch_events":"https://pith.science/api/pith-number/RXAIV4XQ7K6DKJYTDVFDUU7W7M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RXAIV4XQ7K6DKJYTDVFDUU7W7M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RXAIV4XQ7K6DKJYTDVFDUU7W7M/action/storage_attestation","attest_author":"https://pith.science/pith/RXAIV4XQ7K6DKJYTDVFDUU7W7M/action/author_attestation","sign_citation":"https://pith.science/pith/RXAIV4XQ7K6DKJYTDVFDUU7W7M/action/citation_signature","submit_replication":"https://pith.science/pith/RXAIV4XQ7K6DKJYTDVFDUU7W7M/action/replication_record"}},"created_at":"2026-06-02T02:04:34.923779+00:00","updated_at":"2026-06-02T02:04:34.923779+00:00"}