{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UPKI3B3QSINK3BU2YCH3GUOIGE","short_pith_number":"pith:UPKI3B3Q","schema_version":"1.0","canonical_sha256":"a3d48d8770921aad869ac08fb351c8313f219e346a011079d1ab981806957425","source":{"kind":"arxiv","id":"2607.08215","version":1},"attestation_state":"computed","paper":{"title":"On the Limitations of Non-GPU AI Accelerators for Large-Model Inference: A Field Study of MoE and Multimodal Serving on Huawei Ascend","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Zheng Yu","submitted_at":"2026-07-09T08:12:15Z","abstract_excerpt":"Non-GPU AI accelerators are increasingly adopted as alternatives to general-purpose GPUs for large-model inference, but the real engineering cost of migrating demanding workloads beyond CUDA remains poorly documented. We present a field study of deploying two large inference workloads on a 16-device Huawei Ascend 910 system using CANN and vLLM-Ascend: an LLM-as-a-judge safety and alignment evaluation pipeline based on a W8A8 MoE judge model, DeepSeek-V4-Flash, and a multimodal medical vision--language benchmark based on DeepSeek-V4-Flash-Vision for MMMU and MMMU-Pro. Making these workloads rel"},"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":"2607.08215","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2026-07-09T08:12:15Z","cross_cats_sorted":[],"title_canon_sha256":"7904174c2924f90b42b24827bf1e75bd302fc87654211d6ab301598be2ceff73","abstract_canon_sha256":"d7d555c0b12b2b354c8581d0cd3ce3ccd73faee9deec1fd5076cf95af8d6bffa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-10T01:19:29.919569Z","signature_b64":"itaL88xU1G9Ow6DzgRCVx3kDgVZs/2mHTj6p3YsEA3FaazpCJoYssPeubbrdoPQ4SMetQQ+ZSQH8rrDmHruHCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a3d48d8770921aad869ac08fb351c8313f219e346a011079d1ab981806957425","last_reissued_at":"2026-07-10T01:19:29.919175Z","signature_status":"signed_v1","first_computed_at":"2026-07-10T01:19:29.919175Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Limitations of Non-GPU AI Accelerators for Large-Model Inference: A Field Study of MoE and Multimodal Serving on Huawei Ascend","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Zheng Yu","submitted_at":"2026-07-09T08:12:15Z","abstract_excerpt":"Non-GPU AI accelerators are increasingly adopted as alternatives to general-purpose GPUs for large-model inference, but the real engineering cost of migrating demanding workloads beyond CUDA remains poorly documented. We present a field study of deploying two large inference workloads on a 16-device Huawei Ascend 910 system using CANN and vLLM-Ascend: an LLM-as-a-judge safety and alignment evaluation pipeline based on a W8A8 MoE judge model, DeepSeek-V4-Flash, and a multimodal medical vision--language benchmark based on DeepSeek-V4-Flash-Vision for MMMU and MMMU-Pro. Making these workloads rel"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.08215","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/2607.08215/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":"2607.08215","created_at":"2026-07-10T01:19:29.919240+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.08215v1","created_at":"2026-07-10T01:19:29.919240+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.08215","created_at":"2026-07-10T01:19:29.919240+00:00"},{"alias_kind":"pith_short_12","alias_value":"UPKI3B3QSINK","created_at":"2026-07-10T01:19:29.919240+00:00"},{"alias_kind":"pith_short_16","alias_value":"UPKI3B3QSINK3BU2","created_at":"2026-07-10T01:19:29.919240+00:00"},{"alias_kind":"pith_short_8","alias_value":"UPKI3B3Q","created_at":"2026-07-10T01:19:29.919240+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/UPKI3B3QSINK3BU2YCH3GUOIGE","json":"https://pith.science/pith/UPKI3B3QSINK3BU2YCH3GUOIGE.json","graph_json":"https://pith.science/api/pith-number/UPKI3B3QSINK3BU2YCH3GUOIGE/graph.json","events_json":"https://pith.science/api/pith-number/UPKI3B3QSINK3BU2YCH3GUOIGE/events.json","paper":"https://pith.science/paper/UPKI3B3Q"},"agent_actions":{"view_html":"https://pith.science/pith/UPKI3B3QSINK3BU2YCH3GUOIGE","download_json":"https://pith.science/pith/UPKI3B3QSINK3BU2YCH3GUOIGE.json","view_paper":"https://pith.science/paper/UPKI3B3Q","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.08215&json=true","fetch_graph":"https://pith.science/api/pith-number/UPKI3B3QSINK3BU2YCH3GUOIGE/graph.json","fetch_events":"https://pith.science/api/pith-number/UPKI3B3QSINK3BU2YCH3GUOIGE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UPKI3B3QSINK3BU2YCH3GUOIGE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UPKI3B3QSINK3BU2YCH3GUOIGE/action/storage_attestation","attest_author":"https://pith.science/pith/UPKI3B3QSINK3BU2YCH3GUOIGE/action/author_attestation","sign_citation":"https://pith.science/pith/UPKI3B3QSINK3BU2YCH3GUOIGE/action/citation_signature","submit_replication":"https://pith.science/pith/UPKI3B3QSINK3BU2YCH3GUOIGE/action/replication_record"}},"created_at":"2026-07-10T01:19:29.919240+00:00","updated_at":"2026-07-10T01:19:29.919240+00:00"}