{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:A5IZOGZYAKTPTBOUROQG5VT55O","short_pith_number":"pith:A5IZOGZY","schema_version":"1.0","canonical_sha256":"0751971b3802a6f985d48ba06ed67deb9ba13acac66e6c1c0d988d9d4eb1c06b","source":{"kind":"arxiv","id":"2606.23001","version":1},"attestation_state":"computed","paper":{"title":"EnerInfer: Energy-Aware On-Device LLM Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Binqi Sun, Bohua Zou, Debayan Roy, Haibo Chen, Matteo Mascherin, Nian Liu, Ning Jia, Yu Peng, Yutao Liu","submitted_at":"2026-06-22T08:16:19Z","abstract_excerpt":"On-device LLM inference is increasingly attractive for privacy-preserving, reliable, and cost-effective deployment, yet its energy and thermal costs remain a critical bottleneck. Existing systems primarily optimize for decoding speed, implicitly assuming that faster execution is always preferable. We show instead that on-device LLM inference often has exploitable configuration slack: modestly lowering NPU and memory frequencies preserves quality of experience (QoE) while substantially improving energy efficiency and reducing heat.\n  Realizing this opportunity in production is challenging. The "},"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":"2606.23001","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2026-06-22T08:16:19Z","cross_cats_sorted":[],"title_canon_sha256":"be00aed2a092864ee486b50514cf0a64c8fb4b858b05fdb214dafe4fe42cb415","abstract_canon_sha256":"fb2c7d67248b9abf58cc14e856faccf84b37b804521829352ee7f678da2bab9d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T03:14:06.433898Z","signature_b64":"1m0NeAcHTT3UP6cB7yRaKN8kZ0J/6ljJ9qX6/aB7g0oJeisXRcg+a/Mf4EYNLZEZ4XbtYPN2XvgRISv5fILTAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0751971b3802a6f985d48ba06ed67deb9ba13acac66e6c1c0d988d9d4eb1c06b","last_reissued_at":"2026-06-23T03:14:06.433462Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T03:14:06.433462Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EnerInfer: Energy-Aware On-Device LLM Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Binqi Sun, Bohua Zou, Debayan Roy, Haibo Chen, Matteo Mascherin, Nian Liu, Ning Jia, Yu Peng, Yutao Liu","submitted_at":"2026-06-22T08:16:19Z","abstract_excerpt":"On-device LLM inference is increasingly attractive for privacy-preserving, reliable, and cost-effective deployment, yet its energy and thermal costs remain a critical bottleneck. Existing systems primarily optimize for decoding speed, implicitly assuming that faster execution is always preferable. We show instead that on-device LLM inference often has exploitable configuration slack: modestly lowering NPU and memory frequencies preserves quality of experience (QoE) while substantially improving energy efficiency and reducing heat.\n  Realizing this opportunity in production is challenging. The "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.23001","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/2606.23001/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":"2606.23001","created_at":"2026-06-23T03:14:06.433568+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.23001v1","created_at":"2026-06-23T03:14:06.433568+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.23001","created_at":"2026-06-23T03:14:06.433568+00:00"},{"alias_kind":"pith_short_12","alias_value":"A5IZOGZYAKTP","created_at":"2026-06-23T03:14:06.433568+00:00"},{"alias_kind":"pith_short_16","alias_value":"A5IZOGZYAKTPTBOU","created_at":"2026-06-23T03:14:06.433568+00:00"},{"alias_kind":"pith_short_8","alias_value":"A5IZOGZY","created_at":"2026-06-23T03:14:06.433568+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/A5IZOGZYAKTPTBOUROQG5VT55O","json":"https://pith.science/pith/A5IZOGZYAKTPTBOUROQG5VT55O.json","graph_json":"https://pith.science/api/pith-number/A5IZOGZYAKTPTBOUROQG5VT55O/graph.json","events_json":"https://pith.science/api/pith-number/A5IZOGZYAKTPTBOUROQG5VT55O/events.json","paper":"https://pith.science/paper/A5IZOGZY"},"agent_actions":{"view_html":"https://pith.science/pith/A5IZOGZYAKTPTBOUROQG5VT55O","download_json":"https://pith.science/pith/A5IZOGZYAKTPTBOUROQG5VT55O.json","view_paper":"https://pith.science/paper/A5IZOGZY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.23001&json=true","fetch_graph":"https://pith.science/api/pith-number/A5IZOGZYAKTPTBOUROQG5VT55O/graph.json","fetch_events":"https://pith.science/api/pith-number/A5IZOGZYAKTPTBOUROQG5VT55O/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/A5IZOGZYAKTPTBOUROQG5VT55O/action/timestamp_anchor","attest_storage":"https://pith.science/pith/A5IZOGZYAKTPTBOUROQG5VT55O/action/storage_attestation","attest_author":"https://pith.science/pith/A5IZOGZYAKTPTBOUROQG5VT55O/action/author_attestation","sign_citation":"https://pith.science/pith/A5IZOGZYAKTPTBOUROQG5VT55O/action/citation_signature","submit_replication":"https://pith.science/pith/A5IZOGZYAKTPTBOUROQG5VT55O/action/replication_record"}},"created_at":"2026-06-23T03:14:06.433568+00:00","updated_at":"2026-06-23T03:14:06.433568+00:00"}