{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FABQALSQMYDXXMWJXKCFWJOQT7","short_pith_number":"pith:FABQALSQ","schema_version":"1.0","canonical_sha256":"2803002e5066077bb2c9ba845b25d09fcba46836d8bbef1ea2c80fc447c7b8ca","source":{"kind":"arxiv","id":"2608.01785","version":1},"attestation_state":"computed","paper":{"title":"HorizonServe: Coordinating Request Scheduling with GPU Sharing for Omni-Model Serving","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Dong Yuan, Yuning Zhang","submitted_at":"2026-08-03T06:59:15Z","abstract_excerpt":"Omni models unify text, speech, image, and multimodal reasoning in a single serving backend, but this unified deployment exposes a new scheduling problem. Requests with different output modalities may share an initial multimodal backbone and then diverge into downstream generation stages, creating heterogeneous first-response metrics and service-level objective (SLO) targets on the same GPU. Existing large language model (LLM) and multimodal serving systems mainly optimize token progress or input-side processing, and they do not jointly control temporal sharing in the shared stage and spatial "},"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":"2608.01785","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.DC","submitted_at":"2026-08-03T06:59:15Z","cross_cats_sorted":[],"title_canon_sha256":"c02ded62b48fa1abaa9e3cebc2a1d3b6a19f6014f933368a5055cfedf11793a5","abstract_canon_sha256":"2e4de932ccfd37375aa7ed63cad4ea2baf3bccf2dc0f7a1aa7a9d0c79a8d9eec"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-04T02:07:05.614327Z","signature_b64":"gF9C3yTWjBAOPh/TpG2fuk5UqWQrG4WRfB0ynqznTo0L+Dr5+6wkN2vLjNvF91L0Z8Vr6hESSKzDVcVF950eBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2803002e5066077bb2c9ba845b25d09fcba46836d8bbef1ea2c80fc447c7b8ca","last_reissued_at":"2026-08-04T02:07:05.612867Z","signature_status":"signed_v1","first_computed_at":"2026-08-04T02:07:05.612867Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HorizonServe: Coordinating Request Scheduling with GPU Sharing for Omni-Model Serving","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Dong Yuan, Yuning Zhang","submitted_at":"2026-08-03T06:59:15Z","abstract_excerpt":"Omni models unify text, speech, image, and multimodal reasoning in a single serving backend, but this unified deployment exposes a new scheduling problem. Requests with different output modalities may share an initial multimodal backbone and then diverge into downstream generation stages, creating heterogeneous first-response metrics and service-level objective (SLO) targets on the same GPU. Existing large language model (LLM) and multimodal serving systems mainly optimize token progress or input-side processing, and they do not jointly control temporal sharing in the shared stage and spatial "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.01785","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/2608.01785/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":"2608.01785","created_at":"2026-08-04T02:07:05.614318+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.01785v1","created_at":"2026-08-04T02:07:05.614318+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.01785","created_at":"2026-08-04T02:07:05.614318+00:00"},{"alias_kind":"pith_short_12","alias_value":"FABQALSQMYDX","created_at":"2026-08-04T02:07:05.614318+00:00"},{"alias_kind":"pith_short_16","alias_value":"FABQALSQMYDXXMWJ","created_at":"2026-08-04T02:07:05.614318+00:00"},{"alias_kind":"pith_short_8","alias_value":"FABQALSQ","created_at":"2026-08-04T02:07:05.614318+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/FABQALSQMYDXXMWJXKCFWJOQT7","json":"https://pith.science/pith/FABQALSQMYDXXMWJXKCFWJOQT7.json","graph_json":"https://pith.science/api/pith-number/FABQALSQMYDXXMWJXKCFWJOQT7/graph.json","events_json":"https://pith.science/api/pith-number/FABQALSQMYDXXMWJXKCFWJOQT7/events.json","paper":"https://pith.science/paper/FABQALSQ"},"agent_actions":{"view_html":"https://pith.science/pith/FABQALSQMYDXXMWJXKCFWJOQT7","download_json":"https://pith.science/pith/FABQALSQMYDXXMWJXKCFWJOQT7.json","view_paper":"https://pith.science/paper/FABQALSQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.01785&json=true","fetch_graph":"https://pith.science/api/pith-number/FABQALSQMYDXXMWJXKCFWJOQT7/graph.json","fetch_events":"https://pith.science/api/pith-number/FABQALSQMYDXXMWJXKCFWJOQT7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FABQALSQMYDXXMWJXKCFWJOQT7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FABQALSQMYDXXMWJXKCFWJOQT7/action/storage_attestation","attest_author":"https://pith.science/pith/FABQALSQMYDXXMWJXKCFWJOQT7/action/author_attestation","sign_citation":"https://pith.science/pith/FABQALSQMYDXXMWJXKCFWJOQT7/action/citation_signature","submit_replication":"https://pith.science/pith/FABQALSQMYDXXMWJXKCFWJOQT7/action/replication_record"}},"created_at":"2026-08-04T02:07:05.614318+00:00","updated_at":"2026-08-04T02:07:05.614318+00:00"}