{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:OETDQIPX4DZU6H66IJOBJNLUYR","short_pith_number":"pith:OETDQIPX","schema_version":"1.0","canonical_sha256":"71263821f7e0f34f1fde425c14b574c4757cfc29cbe4a9b087d4a6d4dd1e22d7","source":{"kind":"arxiv","id":"2503.16731","version":3},"attestation_state":"computed","paper":{"title":"Design and Implementation of an FPGA-Based Hardware Accelerator for Transformer","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AR","authors_text":"Richie Li, Sicheng Chen","submitted_at":"2025-03-20T22:15:42Z","abstract_excerpt":"Transformer-based large language models (LLMs) rely heavily on intensive matrix multiplications for attention and feed-forward layers, with the Q, K, and V linear projections in the Multi-Head Self-Attention (MHA) module constituting a decisive performance bottleneck. In this work, we introduce a highly optimized tiled matrix multiplication accelerator on a resource-constrained Xilinx KV260 FPGA that not only addresses this challenge but sets a new standard for efficiency and performance. Our design exploits persistent on-chip storage, a robust two-level tiling strategy for maximal data reuse,"},"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":"2503.16731","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AR","submitted_at":"2025-03-20T22:15:42Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"ba63fc581269d20419a67bfe7ce9d0808f50b432e7774fd5844e235d2e56452a","abstract_canon_sha256":"7e48df9cfa597997b65bc27a7903ee1c8f37ff895f4ef633643a082d640e4658"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:06:29.648125Z","signature_b64":"mFrizPo2KbDkErwb+GkvIaAmraB7LpY9avDBCHyroyM+nP56BMO39OrvWzgFx2L8uQtrwCn7rbs9empDZvymDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"71263821f7e0f34f1fde425c14b574c4757cfc29cbe4a9b087d4a6d4dd1e22d7","last_reissued_at":"2026-07-05T11:06:29.647614Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:06:29.647614Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Design and Implementation of an FPGA-Based Hardware Accelerator for Transformer","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AR","authors_text":"Richie Li, Sicheng Chen","submitted_at":"2025-03-20T22:15:42Z","abstract_excerpt":"Transformer-based large language models (LLMs) rely heavily on intensive matrix multiplications for attention and feed-forward layers, with the Q, K, and V linear projections in the Multi-Head Self-Attention (MHA) module constituting a decisive performance bottleneck. In this work, we introduce a highly optimized tiled matrix multiplication accelerator on a resource-constrained Xilinx KV260 FPGA that not only addresses this challenge but sets a new standard for efficiency and performance. Our design exploits persistent on-chip storage, a robust two-level tiling strategy for maximal data reuse,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.16731","kind":"arxiv","version":3},"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/2503.16731/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":"2503.16731","created_at":"2026-07-05T11:06:29.647675+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.16731v3","created_at":"2026-07-05T11:06:29.647675+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.16731","created_at":"2026-07-05T11:06:29.647675+00:00"},{"alias_kind":"pith_short_12","alias_value":"OETDQIPX4DZU","created_at":"2026-07-05T11:06:29.647675+00:00"},{"alias_kind":"pith_short_16","alias_value":"OETDQIPX4DZU6H66","created_at":"2026-07-05T11:06:29.647675+00:00"},{"alias_kind":"pith_short_8","alias_value":"OETDQIPX","created_at":"2026-07-05T11:06:29.647675+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07713","citing_title":"Attention at the Theoretical Minimum: A Mathematics of Arrays Framework for Memory-Optimal Transformer Kernels","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2602.12851","citing_title":"Chimera: Neuro-Symbolic Attention Primitives for Trustworthy Dataplane Intelligence","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OETDQIPX4DZU6H66IJOBJNLUYR","json":"https://pith.science/pith/OETDQIPX4DZU6H66IJOBJNLUYR.json","graph_json":"https://pith.science/api/pith-number/OETDQIPX4DZU6H66IJOBJNLUYR/graph.json","events_json":"https://pith.science/api/pith-number/OETDQIPX4DZU6H66IJOBJNLUYR/events.json","paper":"https://pith.science/paper/OETDQIPX"},"agent_actions":{"view_html":"https://pith.science/pith/OETDQIPX4DZU6H66IJOBJNLUYR","download_json":"https://pith.science/pith/OETDQIPX4DZU6H66IJOBJNLUYR.json","view_paper":"https://pith.science/paper/OETDQIPX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.16731&json=true","fetch_graph":"https://pith.science/api/pith-number/OETDQIPX4DZU6H66IJOBJNLUYR/graph.json","fetch_events":"https://pith.science/api/pith-number/OETDQIPX4DZU6H66IJOBJNLUYR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OETDQIPX4DZU6H66IJOBJNLUYR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OETDQIPX4DZU6H66IJOBJNLUYR/action/storage_attestation","attest_author":"https://pith.science/pith/OETDQIPX4DZU6H66IJOBJNLUYR/action/author_attestation","sign_citation":"https://pith.science/pith/OETDQIPX4DZU6H66IJOBJNLUYR/action/citation_signature","submit_replication":"https://pith.science/pith/OETDQIPX4DZU6H66IJOBJNLUYR/action/replication_record"}},"created_at":"2026-07-05T11:06:29.647675+00:00","updated_at":"2026-07-05T11:06:29.647675+00:00"}