{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:LE76IOTVK24PSAHJ4AOZCJMNZW","short_pith_number":"pith:LE76IOTV","schema_version":"1.0","canonical_sha256":"593fe43a7556b8f900e9e01d91258dcda39c0d8b6481e6c588143038860cf2e2","source":{"kind":"arxiv","id":"2407.21325","version":2},"attestation_state":"computed","paper":{"title":"EdgeLLM: A Highly Efficient CPU-FPGA Heterogeneous Edge Accelerator for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Ao Shen, Boyu Li, Haoxiang Peng, Hao Yu, Kai Li, Mingqiang Huang, Yupeng Su","submitted_at":"2024-07-31T04:16:37Z","abstract_excerpt":"The rapid advancements in artificial intelligence (AI), particularly the Large Language Models (LLMs), have profoundly affected our daily work and communication forms. However, it is still a challenge to deploy LLMs on resource-constrained edge devices (such as robots), due to the intensive computation requirements, heavy memory access, diverse operator types and difficulties in compilation. In this work, we proposed EdgeLLM to address the above issues. Firstly, focusing on the computation, we designed mix-precision processing element array together with group systolic architecture, that can e"},"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":"2407.21325","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2024-07-31T04:16:37Z","cross_cats_sorted":[],"title_canon_sha256":"509804d4e0706f49dbb692aa6d0b3f9aad9472f4fd7c37904b2c923ed187d3c6","abstract_canon_sha256":"bdd7b40a56fd919d133e1c77ffc8f704846650d9844efc802ef882230e213970"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:21:22.184442Z","signature_b64":"PJNfmUH/LtEksgYTecvPMbPV+mPj75YbEAbna352tCkNfUOD//YMoj+0CW0+DrGHZM9mr4V5WTB0Xl2RHKD/CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"593fe43a7556b8f900e9e01d91258dcda39c0d8b6481e6c588143038860cf2e2","last_reissued_at":"2026-07-05T10:21:22.183917Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:21:22.183917Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EdgeLLM: A Highly Efficient CPU-FPGA Heterogeneous Edge Accelerator for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Ao Shen, Boyu Li, Haoxiang Peng, Hao Yu, Kai Li, Mingqiang Huang, Yupeng Su","submitted_at":"2024-07-31T04:16:37Z","abstract_excerpt":"The rapid advancements in artificial intelligence (AI), particularly the Large Language Models (LLMs), have profoundly affected our daily work and communication forms. However, it is still a challenge to deploy LLMs on resource-constrained edge devices (such as robots), due to the intensive computation requirements, heavy memory access, diverse operator types and difficulties in compilation. In this work, we proposed EdgeLLM to address the above issues. Firstly, focusing on the computation, we designed mix-precision processing element array together with group systolic architecture, that can e"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.21325","kind":"arxiv","version":2},"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/2407.21325/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":"2407.21325","created_at":"2026-07-05T10:21:22.183977+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.21325v2","created_at":"2026-07-05T10:21:22.183977+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.21325","created_at":"2026-07-05T10:21:22.183977+00:00"},{"alias_kind":"pith_short_12","alias_value":"LE76IOTVK24P","created_at":"2026-07-05T10:21:22.183977+00:00"},{"alias_kind":"pith_short_16","alias_value":"LE76IOTVK24PSAHJ","created_at":"2026-07-05T10:21:22.183977+00:00"},{"alias_kind":"pith_short_8","alias_value":"LE76IOTV","created_at":"2026-07-05T10:21:22.183977+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.20187","citing_title":"Breaking the Boundaries of Long-Context LLM Inference: Adaptive KV Management on a Single Commodity GPU","ref_index":21,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LE76IOTVK24PSAHJ4AOZCJMNZW","json":"https://pith.science/pith/LE76IOTVK24PSAHJ4AOZCJMNZW.json","graph_json":"https://pith.science/api/pith-number/LE76IOTVK24PSAHJ4AOZCJMNZW/graph.json","events_json":"https://pith.science/api/pith-number/LE76IOTVK24PSAHJ4AOZCJMNZW/events.json","paper":"https://pith.science/paper/LE76IOTV"},"agent_actions":{"view_html":"https://pith.science/pith/LE76IOTVK24PSAHJ4AOZCJMNZW","download_json":"https://pith.science/pith/LE76IOTVK24PSAHJ4AOZCJMNZW.json","view_paper":"https://pith.science/paper/LE76IOTV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.21325&json=true","fetch_graph":"https://pith.science/api/pith-number/LE76IOTVK24PSAHJ4AOZCJMNZW/graph.json","fetch_events":"https://pith.science/api/pith-number/LE76IOTVK24PSAHJ4AOZCJMNZW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LE76IOTVK24PSAHJ4AOZCJMNZW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LE76IOTVK24PSAHJ4AOZCJMNZW/action/storage_attestation","attest_author":"https://pith.science/pith/LE76IOTVK24PSAHJ4AOZCJMNZW/action/author_attestation","sign_citation":"https://pith.science/pith/LE76IOTVK24PSAHJ4AOZCJMNZW/action/citation_signature","submit_replication":"https://pith.science/pith/LE76IOTVK24PSAHJ4AOZCJMNZW/action/replication_record"}},"created_at":"2026-07-05T10:21:22.183977+00:00","updated_at":"2026-07-05T10:21:22.183977+00:00"}