{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7TTAGZHYDHT5LO5KUFVDRCFTWW","short_pith_number":"pith:7TTAGZHY","schema_version":"1.0","canonical_sha256":"fce60364f819e7d5bbaaa16a3888b3b5a7ae3e6cd3883676a601191598784e33","source":{"kind":"arxiv","id":"2505.18824","version":1},"attestation_state":"computed","paper":{"title":"FlatAttention: Dataflow and Fabric Collectives Co-Optimization for Efficient Multi-Head Attention on Tile-Based Many-PE Accelerators","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Alessandro Nadalini, Chi Zhang, Francesco Conti, Gamze Islamoglu, Luca Benini, Luca Colagrande, Renzo Andri, Thomas Benz, Yawei Li","submitted_at":"2025-05-24T18:50:04Z","abstract_excerpt":"Multi-Head Attention (MHA) is a critical computational kernel in transformer-based AI models. Emerging scalable tile-based accelerator architectures integrate increasing numbers of tightly-packed processing elements (PEs) with tensor units. MHA dataflow mapping is crucial for achieving high utilization of the available units. We propose FlatAttention, a new dataflow for MHA on tile-based many-PE accelerators, minimizing costly main memory (HBM) accesses by leveraging collective primitives integrated into the on-chip network fabric. FlatAttention achieves up to 89.3% utilization, and 4.1x perfo"},"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":"2505.18824","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AR","submitted_at":"2025-05-24T18:50:04Z","cross_cats_sorted":[],"title_canon_sha256":"ba6f0b656bbd7f00954a24d1722b6082045ee917ac661aa426d3ab8b6aeceb4e","abstract_canon_sha256":"0584729c648e2047ceee11410a53149142cac9f64f966ba60422a5a94fe7f73c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:09:20.171025Z","signature_b64":"UbRE/sq8vlN5djklwCOiZOHk7iZ796neV4MonvzaCggZiacXQlt9tswXn5XQSiacVuGm0I+TgnE3JVZx7uYtBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fce60364f819e7d5bbaaa16a3888b3b5a7ae3e6cd3883676a601191598784e33","last_reissued_at":"2026-07-05T11:09:20.170536Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:09:20.170536Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FlatAttention: Dataflow and Fabric Collectives Co-Optimization for Efficient Multi-Head Attention on Tile-Based Many-PE Accelerators","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Alessandro Nadalini, Chi Zhang, Francesco Conti, Gamze Islamoglu, Luca Benini, Luca Colagrande, Renzo Andri, Thomas Benz, Yawei Li","submitted_at":"2025-05-24T18:50:04Z","abstract_excerpt":"Multi-Head Attention (MHA) is a critical computational kernel in transformer-based AI models. Emerging scalable tile-based accelerator architectures integrate increasing numbers of tightly-packed processing elements (PEs) with tensor units. MHA dataflow mapping is crucial for achieving high utilization of the available units. We propose FlatAttention, a new dataflow for MHA on tile-based many-PE accelerators, minimizing costly main memory (HBM) accesses by leveraging collective primitives integrated into the on-chip network fabric. FlatAttention achieves up to 89.3% utilization, and 4.1x perfo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.18824","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/2505.18824/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":"2505.18824","created_at":"2026-07-05T11:09:20.170598+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.18824v1","created_at":"2026-07-05T11:09:20.170598+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.18824","created_at":"2026-07-05T11:09:20.170598+00:00"},{"alias_kind":"pith_short_12","alias_value":"7TTAGZHYDHT5","created_at":"2026-07-05T11:09:20.170598+00:00"},{"alias_kind":"pith_short_16","alias_value":"7TTAGZHYDHT5LO5K","created_at":"2026-07-05T11:09:20.170598+00:00"},{"alias_kind":"pith_short_8","alias_value":"7TTAGZHY","created_at":"2026-07-05T11:09:20.170598+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/7TTAGZHYDHT5LO5KUFVDRCFTWW","json":"https://pith.science/pith/7TTAGZHYDHT5LO5KUFVDRCFTWW.json","graph_json":"https://pith.science/api/pith-number/7TTAGZHYDHT5LO5KUFVDRCFTWW/graph.json","events_json":"https://pith.science/api/pith-number/7TTAGZHYDHT5LO5KUFVDRCFTWW/events.json","paper":"https://pith.science/paper/7TTAGZHY"},"agent_actions":{"view_html":"https://pith.science/pith/7TTAGZHYDHT5LO5KUFVDRCFTWW","download_json":"https://pith.science/pith/7TTAGZHYDHT5LO5KUFVDRCFTWW.json","view_paper":"https://pith.science/paper/7TTAGZHY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.18824&json=true","fetch_graph":"https://pith.science/api/pith-number/7TTAGZHYDHT5LO5KUFVDRCFTWW/graph.json","fetch_events":"https://pith.science/api/pith-number/7TTAGZHYDHT5LO5KUFVDRCFTWW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7TTAGZHYDHT5LO5KUFVDRCFTWW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7TTAGZHYDHT5LO5KUFVDRCFTWW/action/storage_attestation","attest_author":"https://pith.science/pith/7TTAGZHYDHT5LO5KUFVDRCFTWW/action/author_attestation","sign_citation":"https://pith.science/pith/7TTAGZHYDHT5LO5KUFVDRCFTWW/action/citation_signature","submit_replication":"https://pith.science/pith/7TTAGZHYDHT5LO5KUFVDRCFTWW/action/replication_record"}},"created_at":"2026-07-05T11:09:20.170598+00:00","updated_at":"2026-07-05T11:09:20.170598+00:00"}