{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:IATIAHY3ZT7WSTRCQ37ZT3HPAS","short_pith_number":"pith:IATIAHY3","schema_version":"1.0","canonical_sha256":"4026801f1bccff694e2286ff99ecef0484328c9fe2152df6fc711865a781fd68","source":{"kind":"arxiv","id":"2607.01817","version":1},"attestation_state":"computed","paper":{"title":"HCMS: Head-Chunked Multi-Stream Pipeline for Communication-Computation Overlap in Long-Sequence Parallel Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Chao Yuan, Jing Liu, Pan Li, Yingnan Sun","submitted_at":"2026-07-02T07:30:53Z","abstract_excerpt":"All-to-all based sequence parallelism methods execute communication and computation strictly in serial when processing medium-long sequences, resulting in hardware resource underutilization. This paper proposes Head-Chunked Multi-Stream Pipeline (HCMS), which exploits the computational independence of multi-head attention by partitioning attention heads into multiple chunks and achieving fine-grained communication-computation overlap through dual CUDA streams. HCMS is orthogonally compatible with existing optimizations such as FlashAttention and SDPA, requires no modification to underlying ker"},"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":"2607.01817","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2026-07-02T07:30:53Z","cross_cats_sorted":[],"title_canon_sha256":"c91ccc92007e88ff56cd1417279eb4e650454051992342bfb4afb1e35aac84d7","abstract_canon_sha256":"ec83defdb061cacf7dabbd732ece7bdb9da571100577f11590d845ea625f95dd"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T01:17:30.741275Z","signature_b64":"fBaZNE+fO00+BeqrEZ46pSQxnOD9ECK0/+Zmaj4mAD4hcdM5ikRbpezd9wdfYdUkU6bYwjbGzHFm1BXdwLojCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4026801f1bccff694e2286ff99ecef0484328c9fe2152df6fc711865a781fd68","last_reissued_at":"2026-07-03T01:17:30.740867Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T01:17:30.740867Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HCMS: Head-Chunked Multi-Stream Pipeline for Communication-Computation Overlap in Long-Sequence Parallel Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Chao Yuan, Jing Liu, Pan Li, Yingnan Sun","submitted_at":"2026-07-02T07:30:53Z","abstract_excerpt":"All-to-all based sequence parallelism methods execute communication and computation strictly in serial when processing medium-long sequences, resulting in hardware resource underutilization. This paper proposes Head-Chunked Multi-Stream Pipeline (HCMS), which exploits the computational independence of multi-head attention by partitioning attention heads into multiple chunks and achieving fine-grained communication-computation overlap through dual CUDA streams. HCMS is orthogonally compatible with existing optimizations such as FlashAttention and SDPA, requires no modification to underlying ker"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.01817","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/2607.01817/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":"2607.01817","created_at":"2026-07-03T01:17:30.740920+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.01817v1","created_at":"2026-07-03T01:17:30.740920+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.01817","created_at":"2026-07-03T01:17:30.740920+00:00"},{"alias_kind":"pith_short_12","alias_value":"IATIAHY3ZT7W","created_at":"2026-07-03T01:17:30.740920+00:00"},{"alias_kind":"pith_short_16","alias_value":"IATIAHY3ZT7WSTRC","created_at":"2026-07-03T01:17:30.740920+00:00"},{"alias_kind":"pith_short_8","alias_value":"IATIAHY3","created_at":"2026-07-03T01:17:30.740920+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/IATIAHY3ZT7WSTRCQ37ZT3HPAS","json":"https://pith.science/pith/IATIAHY3ZT7WSTRCQ37ZT3HPAS.json","graph_json":"https://pith.science/api/pith-number/IATIAHY3ZT7WSTRCQ37ZT3HPAS/graph.json","events_json":"https://pith.science/api/pith-number/IATIAHY3ZT7WSTRCQ37ZT3HPAS/events.json","paper":"https://pith.science/paper/IATIAHY3"},"agent_actions":{"view_html":"https://pith.science/pith/IATIAHY3ZT7WSTRCQ37ZT3HPAS","download_json":"https://pith.science/pith/IATIAHY3ZT7WSTRCQ37ZT3HPAS.json","view_paper":"https://pith.science/paper/IATIAHY3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.01817&json=true","fetch_graph":"https://pith.science/api/pith-number/IATIAHY3ZT7WSTRCQ37ZT3HPAS/graph.json","fetch_events":"https://pith.science/api/pith-number/IATIAHY3ZT7WSTRCQ37ZT3HPAS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IATIAHY3ZT7WSTRCQ37ZT3HPAS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IATIAHY3ZT7WSTRCQ37ZT3HPAS/action/storage_attestation","attest_author":"https://pith.science/pith/IATIAHY3ZT7WSTRCQ37ZT3HPAS/action/author_attestation","sign_citation":"https://pith.science/pith/IATIAHY3ZT7WSTRCQ37ZT3HPAS/action/citation_signature","submit_replication":"https://pith.science/pith/IATIAHY3ZT7WSTRCQ37ZT3HPAS/action/replication_record"}},"created_at":"2026-07-03T01:17:30.740920+00:00","updated_at":"2026-07-03T01:17:30.740920+00:00"}