{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:2P5HMBOTO5J3LWMANSCSLZPRUZ","short_pith_number":"pith:2P5HMBOT","schema_version":"1.0","canonical_sha256":"d3fa7605d37753b5d9806c8525e5f1a65c6f28b5f6791f5da03a5a9d5d0a813f","source":{"kind":"arxiv","id":"1708.02188","version":1},"attestation_state":"computed","paper":{"title":"PowerAI DDL","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.DC","authors_text":"David Kung, Dheeraj Sreedhar, Minsik Cho, Sameer Kumar, Ulrich Finkler, Vaibhav Saxena","submitted_at":"2017-08-07T16:24:00Z","abstract_excerpt":"As deep neural networks become more complex and input datasets grow larger, it can take days or even weeks to train a deep neural network to the desired accuracy. Therefore, distributed Deep Learning at a massive scale is a critical capability, since it offers the potential to reduce the training time from weeks to hours. In this paper, we present a software-hardware co-optimized distributed Deep Learning system that can achieve near-linear scaling up to hundreds of GPUs. The core algorithm is a multi-ring communication pattern that provides a good tradeoff between latency and bandwidth and ad"},"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":"1708.02188","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2017-08-07T16:24:00Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"5133eedc0e73d9753be47ac40594db288f8e9daba564943cebd3bb248f2b5ddd","abstract_canon_sha256":"ca4aa07df505f2e11988ff3a626acb3bab5ff0e16212a20326a9502f64efbb12"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:38:31.825573Z","signature_b64":"q7cgrZPfGl8CG6eqJdyzI8/syCbdeZ8d2rn6ssWEp2GBa/UKr/7zClH777nsfcdKicg+Nskhc+N0QGXFr8OkAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d3fa7605d37753b5d9806c8525e5f1a65c6f28b5f6791f5da03a5a9d5d0a813f","last_reissued_at":"2026-05-18T00:38:31.824999Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:38:31.824999Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PowerAI DDL","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.DC","authors_text":"David Kung, Dheeraj Sreedhar, Minsik Cho, Sameer Kumar, Ulrich Finkler, Vaibhav Saxena","submitted_at":"2017-08-07T16:24:00Z","abstract_excerpt":"As deep neural networks become more complex and input datasets grow larger, it can take days or even weeks to train a deep neural network to the desired accuracy. Therefore, distributed Deep Learning at a massive scale is a critical capability, since it offers the potential to reduce the training time from weeks to hours. In this paper, we present a software-hardware co-optimized distributed Deep Learning system that can achieve near-linear scaling up to hundreds of GPUs. The core algorithm is a multi-ring communication pattern that provides a good tradeoff between latency and bandwidth and ad"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1708.02188","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":""},"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":"1708.02188","created_at":"2026-05-18T00:38:31.825105+00:00"},{"alias_kind":"arxiv_version","alias_value":"1708.02188v1","created_at":"2026-05-18T00:38:31.825105+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1708.02188","created_at":"2026-05-18T00:38:31.825105+00:00"},{"alias_kind":"pith_short_12","alias_value":"2P5HMBOTO5J3","created_at":"2026-05-18T12:30:55.937587+00:00"},{"alias_kind":"pith_short_16","alias_value":"2P5HMBOTO5J3LWMA","created_at":"2026-05-18T12:30:55.937587+00:00"},{"alias_kind":"pith_short_8","alias_value":"2P5HMBOT","created_at":"2026-05-18T12:30:55.937587+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"1907.05701","citing_title":"A Highly Efficient Distributed Deep Learning System For Automatic Speech Recognition","ref_index":22,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2P5HMBOTO5J3LWMANSCSLZPRUZ","json":"https://pith.science/pith/2P5HMBOTO5J3LWMANSCSLZPRUZ.json","graph_json":"https://pith.science/api/pith-number/2P5HMBOTO5J3LWMANSCSLZPRUZ/graph.json","events_json":"https://pith.science/api/pith-number/2P5HMBOTO5J3LWMANSCSLZPRUZ/events.json","paper":"https://pith.science/paper/2P5HMBOT"},"agent_actions":{"view_html":"https://pith.science/pith/2P5HMBOTO5J3LWMANSCSLZPRUZ","download_json":"https://pith.science/pith/2P5HMBOTO5J3LWMANSCSLZPRUZ.json","view_paper":"https://pith.science/paper/2P5HMBOT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1708.02188&json=true","fetch_graph":"https://pith.science/api/pith-number/2P5HMBOTO5J3LWMANSCSLZPRUZ/graph.json","fetch_events":"https://pith.science/api/pith-number/2P5HMBOTO5J3LWMANSCSLZPRUZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2P5HMBOTO5J3LWMANSCSLZPRUZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2P5HMBOTO5J3LWMANSCSLZPRUZ/action/storage_attestation","attest_author":"https://pith.science/pith/2P5HMBOTO5J3LWMANSCSLZPRUZ/action/author_attestation","sign_citation":"https://pith.science/pith/2P5HMBOTO5J3LWMANSCSLZPRUZ/action/citation_signature","submit_replication":"https://pith.science/pith/2P5HMBOTO5J3LWMANSCSLZPRUZ/action/replication_record"}},"created_at":"2026-05-18T00:38:31.825105+00:00","updated_at":"2026-05-18T00:38:31.825105+00:00"}