{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:HZGWGY6NNO5XJGDEJPGYLSMY6Z","short_pith_number":"pith:HZGWGY6N","schema_version":"1.0","canonical_sha256":"3e4d6363cd6bbb7498644bcd85c998f646e8a5ddfe8f048c25ed6bf0f76dd562","source":{"kind":"arxiv","id":"2507.03114","version":1},"attestation_state":"computed","paper":{"title":"Characterizing Compute-Communication Overlap in GPU-Accelerated Distributed Deep Learning: Performance and Power Implications","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Divya Mahajan, Jihwan Oh, Jongse Park, Junkyum Kim, Seokjin Go, Seonho Lee","submitted_at":"2025-07-03T18:53:03Z","abstract_excerpt":"This paper provides an in-depth characterization of GPU-accelerated systems, to understand the interplay between overlapping computation and communication which is commonly employed in distributed training settings. Due to the large size of models, distributing them across multiple devices is required. Overlapping strategies, which enable concurrent computation and communication, are critical for mitigating communication bottlenecks and maximizing GPU utilization. However, the current consensus is that we should always and aggressively overlap compute and communication to mitigate the overhead"},"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":"2507.03114","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2025-07-03T18:53:03Z","cross_cats_sorted":[],"title_canon_sha256":"265ffffcf7b27ed5a638355315843e134b7cbd4e352d6df901d9436e70cac375","abstract_canon_sha256":"e2e995231703b9d43fb96bf06693e2263a5ad3ac55de9388e31ad8499b7dae71"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:31:52.814023Z","signature_b64":"WvQLbnXlzRRVDDnmQUCW5ye72eKO/SEhdpsPnXm9+Kv8xqrMG2g1NZ+6JQXM/KMHZ17B6XjxzjwS+X7dlBd9BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3e4d6363cd6bbb7498644bcd85c998f646e8a5ddfe8f048c25ed6bf0f76dd562","last_reissued_at":"2026-07-05T11:31:52.813536Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:31:52.813536Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Characterizing Compute-Communication Overlap in GPU-Accelerated Distributed Deep Learning: Performance and Power Implications","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Divya Mahajan, Jihwan Oh, Jongse Park, Junkyum Kim, Seokjin Go, Seonho Lee","submitted_at":"2025-07-03T18:53:03Z","abstract_excerpt":"This paper provides an in-depth characterization of GPU-accelerated systems, to understand the interplay between overlapping computation and communication which is commonly employed in distributed training settings. Due to the large size of models, distributing them across multiple devices is required. Overlapping strategies, which enable concurrent computation and communication, are critical for mitigating communication bottlenecks and maximizing GPU utilization. However, the current consensus is that we should always and aggressively overlap compute and communication to mitigate the overhead"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.03114","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/2507.03114/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":"2507.03114","created_at":"2026-07-05T11:31:52.813592+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.03114v1","created_at":"2026-07-05T11:31:52.813592+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.03114","created_at":"2026-07-05T11:31:52.813592+00:00"},{"alias_kind":"pith_short_12","alias_value":"HZGWGY6NNO5X","created_at":"2026-07-05T11:31:52.813592+00:00"},{"alias_kind":"pith_short_16","alias_value":"HZGWGY6NNO5XJGDE","created_at":"2026-07-05T11:31:52.813592+00:00"},{"alias_kind":"pith_short_8","alias_value":"HZGWGY6N","created_at":"2026-07-05T11:31:52.813592+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.14249","citing_title":"EnergyLens: Predictive Energy-Aware Exploration for Multi-GPU LLM Inference Optimization","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HZGWGY6NNO5XJGDEJPGYLSMY6Z","json":"https://pith.science/pith/HZGWGY6NNO5XJGDEJPGYLSMY6Z.json","graph_json":"https://pith.science/api/pith-number/HZGWGY6NNO5XJGDEJPGYLSMY6Z/graph.json","events_json":"https://pith.science/api/pith-number/HZGWGY6NNO5XJGDEJPGYLSMY6Z/events.json","paper":"https://pith.science/paper/HZGWGY6N"},"agent_actions":{"view_html":"https://pith.science/pith/HZGWGY6NNO5XJGDEJPGYLSMY6Z","download_json":"https://pith.science/pith/HZGWGY6NNO5XJGDEJPGYLSMY6Z.json","view_paper":"https://pith.science/paper/HZGWGY6N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.03114&json=true","fetch_graph":"https://pith.science/api/pith-number/HZGWGY6NNO5XJGDEJPGYLSMY6Z/graph.json","fetch_events":"https://pith.science/api/pith-number/HZGWGY6NNO5XJGDEJPGYLSMY6Z/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HZGWGY6NNO5XJGDEJPGYLSMY6Z/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HZGWGY6NNO5XJGDEJPGYLSMY6Z/action/storage_attestation","attest_author":"https://pith.science/pith/HZGWGY6NNO5XJGDEJPGYLSMY6Z/action/author_attestation","sign_citation":"https://pith.science/pith/HZGWGY6NNO5XJGDEJPGYLSMY6Z/action/citation_signature","submit_replication":"https://pith.science/pith/HZGWGY6NNO5XJGDEJPGYLSMY6Z/action/replication_record"}},"created_at":"2026-07-05T11:31:52.813592+00:00","updated_at":"2026-07-05T11:31:52.813592+00:00"}