{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:X4QM6J2AFM52MZPUIBEKDSBRWR","short_pith_number":"pith:X4QM6J2A","schema_version":"1.0","canonical_sha256":"bf20cf27402b3ba665f44048a1c831b4710db0a6fca665f3f61e308918b8bd51","source":{"kind":"arxiv","id":"2605.29346","version":1},"attestation_state":"computed","paper":{"title":"Understanding and Reducing Metadata-Driven Host Overheads in Sampling-Based GNN Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Bin Ren, Guannan Wang, Pradeep Kumar, Saima Afrin, Yidong Gong, Yuchen Ma","submitted_at":"2026-05-28T04:37:24Z","abstract_excerpt":"Modern deep learning workloads increasingly exhibit dynamic, metadata-driven execution, where runtime-generated information determines memory provisioning and kernel launch decisions. In sampling-based graph neural network (GNN) training, this behavior places the CPU on the critical path, introducing persistent host-device orchestration overhead and frequent GPU-CPU synchronization, which dominate end-to-end runtime when GPU computation is small. Existing approaches, including CUDA Graphs and GPU dynamic parallelism, fail to address this problem because the metadata-driven control loop remains"},"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":"2605.29346","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2026-05-28T04:37:24Z","cross_cats_sorted":[],"title_canon_sha256":"ea43aa16cb817c3b6a197a8f30de95cb4197e0a98f83f9053d0800eef92894f1","abstract_canon_sha256":"33c82ee98909f1751beb28549de12c94d04c979e32341d7b878665137e50d78c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T01:05:34.227022Z","signature_b64":"7w05f/iAz0BU5YC7GulQj6O9R5j7/uHDWpbec4bt4OVCBYy51Yt0eFtm94qq7vXndBuK0xpzi7t4dOpe2uweBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bf20cf27402b3ba665f44048a1c831b4710db0a6fca665f3f61e308918b8bd51","last_reissued_at":"2026-05-29T01:05:34.226262Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T01:05:34.226262Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Understanding and Reducing Metadata-Driven Host Overheads in Sampling-Based GNN Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Bin Ren, Guannan Wang, Pradeep Kumar, Saima Afrin, Yidong Gong, Yuchen Ma","submitted_at":"2026-05-28T04:37:24Z","abstract_excerpt":"Modern deep learning workloads increasingly exhibit dynamic, metadata-driven execution, where runtime-generated information determines memory provisioning and kernel launch decisions. In sampling-based graph neural network (GNN) training, this behavior places the CPU on the critical path, introducing persistent host-device orchestration overhead and frequent GPU-CPU synchronization, which dominate end-to-end runtime when GPU computation is small. Existing approaches, including CUDA Graphs and GPU dynamic parallelism, fail to address this problem because the metadata-driven control loop remains"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.29346","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/2605.29346/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":"2605.29346","created_at":"2026-05-29T01:05:34.226401+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.29346v1","created_at":"2026-05-29T01:05:34.226401+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.29346","created_at":"2026-05-29T01:05:34.226401+00:00"},{"alias_kind":"pith_short_12","alias_value":"X4QM6J2AFM52","created_at":"2026-05-29T01:05:34.226401+00:00"},{"alias_kind":"pith_short_16","alias_value":"X4QM6J2AFM52MZPU","created_at":"2026-05-29T01:05:34.226401+00:00"},{"alias_kind":"pith_short_8","alias_value":"X4QM6J2A","created_at":"2026-05-29T01:05:34.226401+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/X4QM6J2AFM52MZPUIBEKDSBRWR","json":"https://pith.science/pith/X4QM6J2AFM52MZPUIBEKDSBRWR.json","graph_json":"https://pith.science/api/pith-number/X4QM6J2AFM52MZPUIBEKDSBRWR/graph.json","events_json":"https://pith.science/api/pith-number/X4QM6J2AFM52MZPUIBEKDSBRWR/events.json","paper":"https://pith.science/paper/X4QM6J2A"},"agent_actions":{"view_html":"https://pith.science/pith/X4QM6J2AFM52MZPUIBEKDSBRWR","download_json":"https://pith.science/pith/X4QM6J2AFM52MZPUIBEKDSBRWR.json","view_paper":"https://pith.science/paper/X4QM6J2A","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.29346&json=true","fetch_graph":"https://pith.science/api/pith-number/X4QM6J2AFM52MZPUIBEKDSBRWR/graph.json","fetch_events":"https://pith.science/api/pith-number/X4QM6J2AFM52MZPUIBEKDSBRWR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X4QM6J2AFM52MZPUIBEKDSBRWR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X4QM6J2AFM52MZPUIBEKDSBRWR/action/storage_attestation","attest_author":"https://pith.science/pith/X4QM6J2AFM52MZPUIBEKDSBRWR/action/author_attestation","sign_citation":"https://pith.science/pith/X4QM6J2AFM52MZPUIBEKDSBRWR/action/citation_signature","submit_replication":"https://pith.science/pith/X4QM6J2AFM52MZPUIBEKDSBRWR/action/replication_record"}},"created_at":"2026-05-29T01:05:34.226401+00:00","updated_at":"2026-05-29T01:05:34.226401+00:00"}