{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IRIDTZ3VDEB44EW3C7HWVKVRGU","short_pith_number":"pith:IRIDTZ3V","schema_version":"1.0","canonical_sha256":"445039e7751903ce12db17cf6aaab1352aff3f68f728f1194782b958b0129b8d","source":{"kind":"arxiv","id":"2410.17840","version":2},"attestation_state":"computed","paper":{"title":"Is the GPU Half-Empty or Half-Full? Practical Scheduling Techniques for LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bruce Fontaine, Daya Khudia, Ferdi Kossmann, Michael Cafarella, Samuel Madden","submitted_at":"2024-10-23T13:05:46Z","abstract_excerpt":"Serving systems for Large Language Models (LLMs) improve throughput by processing several requests concurrently. However, multiplexing hardware resources between concurrent requests involves non-trivial scheduling decisions. Practical serving systems typically implement these decisions at two levels: First, a load balancer routes requests to different servers which each hold a replica of the LLM. Then, on each server, an engine-level scheduler decides when to run a request, or when to queue or preempt it. Improved scheduling policies may benefit a wide range of LLM deployments and can often be"},"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":"2410.17840","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-23T13:05:46Z","cross_cats_sorted":[],"title_canon_sha256":"427e63ebc7249b3bc69efe3fbf9bfec46467948c9928e6663e44e7b515b85b10","abstract_canon_sha256":"fab4154fba9b03c6a6d9b4df528d04259737e9a156964edaa6f16bce028290f7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:06:08.735064Z","signature_b64":"4GQg2qrbf5KCi4ifg3Uzd7lhlLfZ8PlIMAXviZja3+lApM+bLzmr0KW5Rvw2UJMyhHyaI/gVCZ5CKjF2b+cmAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"445039e7751903ce12db17cf6aaab1352aff3f68f728f1194782b958b0129b8d","last_reissued_at":"2026-07-05T10:06:08.734548Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:06:08.734548Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Is the GPU Half-Empty or Half-Full? Practical Scheduling Techniques for LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bruce Fontaine, Daya Khudia, Ferdi Kossmann, Michael Cafarella, Samuel Madden","submitted_at":"2024-10-23T13:05:46Z","abstract_excerpt":"Serving systems for Large Language Models (LLMs) improve throughput by processing several requests concurrently. However, multiplexing hardware resources between concurrent requests involves non-trivial scheduling decisions. Practical serving systems typically implement these decisions at two levels: First, a load balancer routes requests to different servers which each hold a replica of the LLM. Then, on each server, an engine-level scheduler decides when to run a request, or when to queue or preempt it. Improved scheduling policies may benefit a wide range of LLM deployments and can often be"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.17840","kind":"arxiv","version":2},"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/2410.17840/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":"2410.17840","created_at":"2026-07-05T10:06:08.734602+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.17840v2","created_at":"2026-07-05T10:06:08.734602+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.17840","created_at":"2026-07-05T10:06:08.734602+00:00"},{"alias_kind":"pith_short_12","alias_value":"IRIDTZ3VDEB4","created_at":"2026-07-05T10:06:08.734602+00:00"},{"alias_kind":"pith_short_16","alias_value":"IRIDTZ3VDEB44EW3","created_at":"2026-07-05T10:06:08.734602+00:00"},{"alias_kind":"pith_short_8","alias_value":"IRIDTZ3V","created_at":"2026-07-05T10:06:08.734602+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/IRIDTZ3VDEB44EW3C7HWVKVRGU","json":"https://pith.science/pith/IRIDTZ3VDEB44EW3C7HWVKVRGU.json","graph_json":"https://pith.science/api/pith-number/IRIDTZ3VDEB44EW3C7HWVKVRGU/graph.json","events_json":"https://pith.science/api/pith-number/IRIDTZ3VDEB44EW3C7HWVKVRGU/events.json","paper":"https://pith.science/paper/IRIDTZ3V"},"agent_actions":{"view_html":"https://pith.science/pith/IRIDTZ3VDEB44EW3C7HWVKVRGU","download_json":"https://pith.science/pith/IRIDTZ3VDEB44EW3C7HWVKVRGU.json","view_paper":"https://pith.science/paper/IRIDTZ3V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.17840&json=true","fetch_graph":"https://pith.science/api/pith-number/IRIDTZ3VDEB44EW3C7HWVKVRGU/graph.json","fetch_events":"https://pith.science/api/pith-number/IRIDTZ3VDEB44EW3C7HWVKVRGU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IRIDTZ3VDEB44EW3C7HWVKVRGU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IRIDTZ3VDEB44EW3C7HWVKVRGU/action/storage_attestation","attest_author":"https://pith.science/pith/IRIDTZ3VDEB44EW3C7HWVKVRGU/action/author_attestation","sign_citation":"https://pith.science/pith/IRIDTZ3VDEB44EW3C7HWVKVRGU/action/citation_signature","submit_replication":"https://pith.science/pith/IRIDTZ3VDEB44EW3C7HWVKVRGU/action/replication_record"}},"created_at":"2026-07-05T10:06:08.734602+00:00","updated_at":"2026-07-05T10:06:08.734602+00:00"}