{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:IDNOLVHFQNTW2L3G44PCMGPHGV","short_pith_number":"pith:IDNOLVHF","schema_version":"1.0","canonical_sha256":"40dae5d4e583676d2f66e71e2619e7355cce91cd9500e419d6570560ab52e204","source":{"kind":"arxiv","id":"2509.09420","version":1},"attestation_state":"computed","paper":{"title":"HD-MoE: Hybrid and Dynamic Parallelism for Mixture-of-Expert LLMs with 3D Near-Memory Processing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.PF","authors_text":"Dimin Niu, Haochen Huang, Hongzhong Zheng, Meng Li, Runsheng Wang, Shuangchen Li, Shuzhang Zhong, Zhe Zhang","submitted_at":"2025-09-11T13:01:55Z","abstract_excerpt":"Large Language Models (LLMs) with Mixture-of-Expert (MoE) architectures achieve superior model performance with reduced computation costs, but at the cost of high memory capacity and bandwidth requirements. Near-Memory Processing (NMP) accelerators that stack memory directly on the compute through hybrid bonding have demonstrated high bandwidth with high energy efficiency, becoming a promising architecture for MoE models. However, as NMP accelerators comprise distributed memory and computation, how to map the MoE computation directly determines the LLM inference efficiency. Existing parallel m"},"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":"2509.09420","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.PF","submitted_at":"2025-09-11T13:01:55Z","cross_cats_sorted":[],"title_canon_sha256":"2fe6ef5b19cd81298e1b84a4ae807a4af0751850590559f22df1e5ba9615ba44","abstract_canon_sha256":"4c46239b9f0a50ff33ec97a8de0cddf8ef6bf888786a5ddd515f16cc91d874c7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:09:35.843998Z","signature_b64":"tL7wCxpKhe8dqeai5mIBS88iZHBwHnpAXaanN+WUxKLiO4BK2d1lPPYWNEUBNIpXq6zZvwlaJxIhPiW74FiGBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"40dae5d4e583676d2f66e71e2619e7355cce91cd9500e419d6570560ab52e204","last_reissued_at":"2026-07-05T12:09:35.843330Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:09:35.843330Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HD-MoE: Hybrid and Dynamic Parallelism for Mixture-of-Expert LLMs with 3D Near-Memory Processing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.PF","authors_text":"Dimin Niu, Haochen Huang, Hongzhong Zheng, Meng Li, Runsheng Wang, Shuangchen Li, Shuzhang Zhong, Zhe Zhang","submitted_at":"2025-09-11T13:01:55Z","abstract_excerpt":"Large Language Models (LLMs) with Mixture-of-Expert (MoE) architectures achieve superior model performance with reduced computation costs, but at the cost of high memory capacity and bandwidth requirements. Near-Memory Processing (NMP) accelerators that stack memory directly on the compute through hybrid bonding have demonstrated high bandwidth with high energy efficiency, becoming a promising architecture for MoE models. However, as NMP accelerators comprise distributed memory and computation, how to map the MoE computation directly determines the LLM inference efficiency. Existing parallel m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.09420","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/2509.09420/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":"2509.09420","created_at":"2026-07-05T12:09:35.843396+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.09420v1","created_at":"2026-07-05T12:09:35.843396+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.09420","created_at":"2026-07-05T12:09:35.843396+00:00"},{"alias_kind":"pith_short_12","alias_value":"IDNOLVHFQNTW","created_at":"2026-07-05T12:09:35.843396+00:00"},{"alias_kind":"pith_short_16","alias_value":"IDNOLVHFQNTW2L3G","created_at":"2026-07-05T12:09:35.843396+00:00"},{"alias_kind":"pith_short_8","alias_value":"IDNOLVHF","created_at":"2026-07-05T12:09:35.843396+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/IDNOLVHFQNTW2L3G44PCMGPHGV","json":"https://pith.science/pith/IDNOLVHFQNTW2L3G44PCMGPHGV.json","graph_json":"https://pith.science/api/pith-number/IDNOLVHFQNTW2L3G44PCMGPHGV/graph.json","events_json":"https://pith.science/api/pith-number/IDNOLVHFQNTW2L3G44PCMGPHGV/events.json","paper":"https://pith.science/paper/IDNOLVHF"},"agent_actions":{"view_html":"https://pith.science/pith/IDNOLVHFQNTW2L3G44PCMGPHGV","download_json":"https://pith.science/pith/IDNOLVHFQNTW2L3G44PCMGPHGV.json","view_paper":"https://pith.science/paper/IDNOLVHF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.09420&json=true","fetch_graph":"https://pith.science/api/pith-number/IDNOLVHFQNTW2L3G44PCMGPHGV/graph.json","fetch_events":"https://pith.science/api/pith-number/IDNOLVHFQNTW2L3G44PCMGPHGV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IDNOLVHFQNTW2L3G44PCMGPHGV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IDNOLVHFQNTW2L3G44PCMGPHGV/action/storage_attestation","attest_author":"https://pith.science/pith/IDNOLVHFQNTW2L3G44PCMGPHGV/action/author_attestation","sign_citation":"https://pith.science/pith/IDNOLVHFQNTW2L3G44PCMGPHGV/action/citation_signature","submit_replication":"https://pith.science/pith/IDNOLVHFQNTW2L3G44PCMGPHGV/action/replication_record"}},"created_at":"2026-07-05T12:09:35.843396+00:00","updated_at":"2026-07-05T12:09:35.843396+00:00"}