{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:Z5ZW7SYMYVHCLS4YXKUFAPFZCD","short_pith_number":"pith:Z5ZW7SYM","schema_version":"1.0","canonical_sha256":"cf736fcb0cc54e25cb98baa8503cb910ffc7d8e4c642a3ca75104d200b23edda","source":{"kind":"arxiv","id":"2605.25655","version":1},"attestation_state":"computed","paper":{"title":"Bandwidth-Aware LLM Inference on Heterogeneous Many-Core Supercomputers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Bin Han, Depei Qian, Gen Li, Hailong Yang, Jiaxing Qi, Shiqing Ma, Shizhe Shang, Yao Lu, Zhongzhi Luan","submitted_at":"2026-05-25T10:03:25Z","abstract_excerpt":"Large language model (LLM) inference is limited by high computational cost and memory bandwidth demands, making deployment on heterogeneous many-core processors challenging. Taking the MT-3000 processor used in the Tianhe supercomputer as an example, its limited main-memory bandwidth and distributed memory hierarchy exemplify these bottlenecks, making it difficult to directly migrate existing GPU-based inference frameworks. To address this problem, we propose THInfer, a hardware-aware inference framework that maximizes data locality under bandwidth-constrained conditions through hardware-softw"},"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":"2605.25655","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2026-05-25T10:03:25Z","cross_cats_sorted":[],"title_canon_sha256":"6651591921947b12e9304c9df84b8e3df11cf8a1843b9f2bbb7d64ee7a65cea0","abstract_canon_sha256":"b3de9869f80eac7eb2d9767a625df360bec9c8903af3cf7ae987a3480d4b841d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:04:48.551978Z","signature_b64":"mpaaYK+GupdiKt55VGgyCPs38EHWiJB1lyKyYZ/PbeqQ+AhZL9hz5i7COnShTT3qczgv1hMOfKkFEqHRZenUAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cf736fcb0cc54e25cb98baa8503cb910ffc7d8e4c642a3ca75104d200b23edda","last_reissued_at":"2026-05-26T02:04:48.550968Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:04:48.550968Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Bandwidth-Aware LLM Inference on Heterogeneous Many-Core Supercomputers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Bin Han, Depei Qian, Gen Li, Hailong Yang, Jiaxing Qi, Shiqing Ma, Shizhe Shang, Yao Lu, Zhongzhi Luan","submitted_at":"2026-05-25T10:03:25Z","abstract_excerpt":"Large language model (LLM) inference is limited by high computational cost and memory bandwidth demands, making deployment on heterogeneous many-core processors challenging. Taking the MT-3000 processor used in the Tianhe supercomputer as an example, its limited main-memory bandwidth and distributed memory hierarchy exemplify these bottlenecks, making it difficult to directly migrate existing GPU-based inference frameworks. To address this problem, we propose THInfer, a hardware-aware inference framework that maximizes data locality under bandwidth-constrained conditions through hardware-softw"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25655","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/2605.25655/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":"2605.25655","created_at":"2026-05-26T02:04:48.551121+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.25655v1","created_at":"2026-05-26T02:04:48.551121+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25655","created_at":"2026-05-26T02:04:48.551121+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z5ZW7SYMYVHC","created_at":"2026-05-26T02:04:48.551121+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z5ZW7SYMYVHCLS4Y","created_at":"2026-05-26T02:04:48.551121+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z5ZW7SYM","created_at":"2026-05-26T02:04:48.551121+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/Z5ZW7SYMYVHCLS4YXKUFAPFZCD","json":"https://pith.science/pith/Z5ZW7SYMYVHCLS4YXKUFAPFZCD.json","graph_json":"https://pith.science/api/pith-number/Z5ZW7SYMYVHCLS4YXKUFAPFZCD/graph.json","events_json":"https://pith.science/api/pith-number/Z5ZW7SYMYVHCLS4YXKUFAPFZCD/events.json","paper":"https://pith.science/paper/Z5ZW7SYM"},"agent_actions":{"view_html":"https://pith.science/pith/Z5ZW7SYMYVHCLS4YXKUFAPFZCD","download_json":"https://pith.science/pith/Z5ZW7SYMYVHCLS4YXKUFAPFZCD.json","view_paper":"https://pith.science/paper/Z5ZW7SYM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.25655&json=true","fetch_graph":"https://pith.science/api/pith-number/Z5ZW7SYMYVHCLS4YXKUFAPFZCD/graph.json","fetch_events":"https://pith.science/api/pith-number/Z5ZW7SYMYVHCLS4YXKUFAPFZCD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z5ZW7SYMYVHCLS4YXKUFAPFZCD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z5ZW7SYMYVHCLS4YXKUFAPFZCD/action/storage_attestation","attest_author":"https://pith.science/pith/Z5ZW7SYMYVHCLS4YXKUFAPFZCD/action/author_attestation","sign_citation":"https://pith.science/pith/Z5ZW7SYMYVHCLS4YXKUFAPFZCD/action/citation_signature","submit_replication":"https://pith.science/pith/Z5ZW7SYMYVHCLS4YXKUFAPFZCD/action/replication_record"}},"created_at":"2026-05-26T02:04:48.551121+00:00","updated_at":"2026-05-26T02:04:48.551121+00:00"}