{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:UUA4CEJ53EH5X36T2DPLYBYZQR","short_pith_number":"pith:UUA4CEJ5","schema_version":"1.0","canonical_sha256":"a501c1113dd90fdbefd3d0debc0719844b5088869a56e30fdcb5679c6be61594","source":{"kind":"arxiv","id":"2408.08586","version":1},"attestation_state":"computed","paper":{"title":"Rubick: Exploiting Job Reconfigurability for Deep Learning Cluster Scheduling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Fangming Liu, Fei Xu, Hanyu Zhao, Wei Lin, Wencong Xiao, Xianyan Jia, Xinyi Zhang, Yong Li","submitted_at":"2024-08-16T07:41:52Z","abstract_excerpt":"The era of large deep learning models has given rise to advanced training strategies such as 3D parallelism and the ZeRO series. These strategies enable various (re-)configurable execution plans for a training job, which exhibit remarkably different requirements of multiple resource types. Existing cluster scheduling systems, however, treat such reconfigurable training jobs as black boxes: they rely on users to choose execution plans statically, and then make resource allocations without awareness of the chosen plans and their resource requirements. This approach results in mismatches between "},"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":"2408.08586","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2024-08-16T07:41:52Z","cross_cats_sorted":[],"title_canon_sha256":"f0ef24d3171a7a1d640b332dc7585cbe1a48f27e0fe1745a5f8d0095606f5d30","abstract_canon_sha256":"27eb34d148b736e88ae39562373744f99f48bc64eabf51404e4a5fc7c5db1556"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:55:57.077022Z","signature_b64":"xX65hyeILe9tCT/X1m3ssG+zzordO6AUeFuJZ8mRUCNtwQxP/Y7+NSf1o7SELjYHsa+dwsYPPLSFKGF3kO93Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a501c1113dd90fdbefd3d0debc0719844b5088869a56e30fdcb5679c6be61594","last_reissued_at":"2026-07-05T08:55:57.076573Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:55:57.076573Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rubick: Exploiting Job Reconfigurability for Deep Learning Cluster Scheduling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Fangming Liu, Fei Xu, Hanyu Zhao, Wei Lin, Wencong Xiao, Xianyan Jia, Xinyi Zhang, Yong Li","submitted_at":"2024-08-16T07:41:52Z","abstract_excerpt":"The era of large deep learning models has given rise to advanced training strategies such as 3D parallelism and the ZeRO series. These strategies enable various (re-)configurable execution plans for a training job, which exhibit remarkably different requirements of multiple resource types. Existing cluster scheduling systems, however, treat such reconfigurable training jobs as black boxes: they rely on users to choose execution plans statically, and then make resource allocations without awareness of the chosen plans and their resource requirements. This approach results in mismatches between "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.08586","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/2408.08586/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":"2408.08586","created_at":"2026-07-05T08:55:57.076639+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.08586v1","created_at":"2026-07-05T08:55:57.076639+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.08586","created_at":"2026-07-05T08:55:57.076639+00:00"},{"alias_kind":"pith_short_12","alias_value":"UUA4CEJ53EH5","created_at":"2026-07-05T08:55:57.076639+00:00"},{"alias_kind":"pith_short_16","alias_value":"UUA4CEJ53EH5X36T","created_at":"2026-07-05T08:55:57.076639+00:00"},{"alias_kind":"pith_short_8","alias_value":"UUA4CEJ5","created_at":"2026-07-05T08:55:57.076639+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.12815","citing_title":"Learning In Chaos: Efficient Autoscaling and Self-Healing for Multi-Party Distributed Training","ref_index":20,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UUA4CEJ53EH5X36T2DPLYBYZQR","json":"https://pith.science/pith/UUA4CEJ53EH5X36T2DPLYBYZQR.json","graph_json":"https://pith.science/api/pith-number/UUA4CEJ53EH5X36T2DPLYBYZQR/graph.json","events_json":"https://pith.science/api/pith-number/UUA4CEJ53EH5X36T2DPLYBYZQR/events.json","paper":"https://pith.science/paper/UUA4CEJ5"},"agent_actions":{"view_html":"https://pith.science/pith/UUA4CEJ53EH5X36T2DPLYBYZQR","download_json":"https://pith.science/pith/UUA4CEJ53EH5X36T2DPLYBYZQR.json","view_paper":"https://pith.science/paper/UUA4CEJ5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.08586&json=true","fetch_graph":"https://pith.science/api/pith-number/UUA4CEJ53EH5X36T2DPLYBYZQR/graph.json","fetch_events":"https://pith.science/api/pith-number/UUA4CEJ53EH5X36T2DPLYBYZQR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UUA4CEJ53EH5X36T2DPLYBYZQR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UUA4CEJ53EH5X36T2DPLYBYZQR/action/storage_attestation","attest_author":"https://pith.science/pith/UUA4CEJ53EH5X36T2DPLYBYZQR/action/author_attestation","sign_citation":"https://pith.science/pith/UUA4CEJ53EH5X36T2DPLYBYZQR/action/citation_signature","submit_replication":"https://pith.science/pith/UUA4CEJ53EH5X36T2DPLYBYZQR/action/replication_record"}},"created_at":"2026-07-05T08:55:57.076639+00:00","updated_at":"2026-07-05T08:55:57.076639+00:00"}