{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NNWRY5JNMIAPBHORJ3BZX56ST7","short_pith_number":"pith:NNWRY5JN","schema_version":"1.0","canonical_sha256":"6b6d1c752d6200f09dd14ec39bf7d29fd7d075a344faef485242fb8cef6fa532","source":{"kind":"arxiv","id":"2605.19481","version":1},"attestation_state":"computed","paper":{"title":"C2CServe: Leveraging NVLink-C2C for Elastic Serverless LLM Serving on MIG","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.OS","authors_text":"Ali Zafar Sadiq, Haiying Shen, Mingye Zhang, Rui Yang, Shutian Luo, Wei Wang, Yue Cheng","submitted_at":"2026-05-19T07:34:08Z","abstract_excerpt":"Modern LLM serving is increasingly serverless in shape: large model catalogs, long-tail invocations, and multi-tenant demand. Existing GPU serving systems face a tradeoff: dedicated-GPU allocation wastes scarce HBM under sparse traffic, while GPU time sharing places model initialization and weight loading on the cold-start path. Spatial GPU sharing such as multi-instance GPU (MIG) provides isolation and accounting, but each slice has too little HBM for modern LLM weights.\n  We observe that high-bandwidth CPU--GPU interconnects, such as NVLink-C2C (C2C) in NVIDIA GH200 and GB200 Superchips, cha"},"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":"2605.19481","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.OS","submitted_at":"2026-05-19T07:34:08Z","cross_cats_sorted":[],"title_canon_sha256":"b3e97021eae5ea634932c7e09890feb27dd8785f11134253a060e288dc3dc553","abstract_canon_sha256":"9d7415d7b6b80cf078fc04649d90ca884b197045ad380ad5e3bb2cf5279928d9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T01:05:47.925869Z","signature_b64":"XBTzuTelwgONwXXSPkD/RaZRn+vRsyidao1qQ94GfL7BwDDV3sktz/AOvH19l800ReR/va4C18kpQ5lw2lY8AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6b6d1c752d6200f09dd14ec39bf7d29fd7d075a344faef485242fb8cef6fa532","last_reissued_at":"2026-05-20T01:05:47.925363Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T01:05:47.925363Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"C2CServe: Leveraging NVLink-C2C for Elastic Serverless LLM Serving on MIG","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.OS","authors_text":"Ali Zafar Sadiq, Haiying Shen, Mingye Zhang, Rui Yang, Shutian Luo, Wei Wang, Yue Cheng","submitted_at":"2026-05-19T07:34:08Z","abstract_excerpt":"Modern LLM serving is increasingly serverless in shape: large model catalogs, long-tail invocations, and multi-tenant demand. Existing GPU serving systems face a tradeoff: dedicated-GPU allocation wastes scarce HBM under sparse traffic, while GPU time sharing places model initialization and weight loading on the cold-start path. Spatial GPU sharing such as multi-instance GPU (MIG) provides isolation and accounting, but each slice has too little HBM for modern LLM weights.\n  We observe that high-bandwidth CPU--GPU interconnects, such as NVLink-C2C (C2C) in NVIDIA GH200 and GB200 Superchips, cha"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.19481","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/2605.19481/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":"2605.19481","created_at":"2026-05-20T01:05:47.925438+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.19481v1","created_at":"2026-05-20T01:05:47.925438+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.19481","created_at":"2026-05-20T01:05:47.925438+00:00"},{"alias_kind":"pith_short_12","alias_value":"NNWRY5JNMIAP","created_at":"2026-05-20T01:05:47.925438+00:00"},{"alias_kind":"pith_short_16","alias_value":"NNWRY5JNMIAPBHOR","created_at":"2026-05-20T01:05:47.925438+00:00"},{"alias_kind":"pith_short_8","alias_value":"NNWRY5JN","created_at":"2026-05-20T01:05:47.925438+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/NNWRY5JNMIAPBHORJ3BZX56ST7","json":"https://pith.science/pith/NNWRY5JNMIAPBHORJ3BZX56ST7.json","graph_json":"https://pith.science/api/pith-number/NNWRY5JNMIAPBHORJ3BZX56ST7/graph.json","events_json":"https://pith.science/api/pith-number/NNWRY5JNMIAPBHORJ3BZX56ST7/events.json","paper":"https://pith.science/paper/NNWRY5JN"},"agent_actions":{"view_html":"https://pith.science/pith/NNWRY5JNMIAPBHORJ3BZX56ST7","download_json":"https://pith.science/pith/NNWRY5JNMIAPBHORJ3BZX56ST7.json","view_paper":"https://pith.science/paper/NNWRY5JN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.19481&json=true","fetch_graph":"https://pith.science/api/pith-number/NNWRY5JNMIAPBHORJ3BZX56ST7/graph.json","fetch_events":"https://pith.science/api/pith-number/NNWRY5JNMIAPBHORJ3BZX56ST7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NNWRY5JNMIAPBHORJ3BZX56ST7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NNWRY5JNMIAPBHORJ3BZX56ST7/action/storage_attestation","attest_author":"https://pith.science/pith/NNWRY5JNMIAPBHORJ3BZX56ST7/action/author_attestation","sign_citation":"https://pith.science/pith/NNWRY5JNMIAPBHORJ3BZX56ST7/action/citation_signature","submit_replication":"https://pith.science/pith/NNWRY5JNMIAPBHORJ3BZX56ST7/action/replication_record"}},"created_at":"2026-05-20T01:05:47.925438+00:00","updated_at":"2026-05-20T01:05:47.925438+00:00"}