{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KN7ZSGHKBRNFE3TREMZC4WNIRZ","short_pith_number":"pith:KN7ZSGHK","schema_version":"1.0","canonical_sha256":"537f9918ea0c5a526e7123322e59a88e73916c40eb5d83ea42ace791d40c1467","source":{"kind":"arxiv","id":"2606.09879","version":1},"attestation_state":"computed","paper":{"title":"Operator Fusion for LLM Inference on the Tensix Architecture","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jie Yu, Ke Li, Lili Liu, Qingbo Wu, Ruian Zhang, Wenzhu Wang","submitted_at":"2026-06-03T02:48:34Z","abstract_excerpt":"This study addresses on-device inference bottlenecks of Transformer models on Tenstorrent's Tensix architecture and proposes an operator fusion strategy that enhances data locality. RMSNorm is fused with matrix multiplication in self-attention and in the FFN, enabling back-to-back execution of memory-bound and compute-bound operators in on-chip SRAM to significantly reduce DRAM reads/writes of intermediate results and scheduling overhead. To support multi-core parallelism, a NoC-based multicast mechanism is leveraged in which row/column master nodes efficiently distribute inputs and weights ac"},"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":"2606.09879","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T02:48:34Z","cross_cats_sorted":[],"title_canon_sha256":"18210d45883210ebc796a0c3162325127dce1834ca9ba11fe6ee40ccd4223d81","abstract_canon_sha256":"a35a28a6f8668bee4a2ba59e889ba2dc97ca817415d8adff19218e5c0f4d7a46"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-10T00:08:30.990790Z","signature_b64":"NzbYyu7X47NokMy1tV1ilfmSpc1CzIW+HOaejtPkVFPHlWtS69N0JViSFg702ksoHFqIJw/iJ380Ihoku69MCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"537f9918ea0c5a526e7123322e59a88e73916c40eb5d83ea42ace791d40c1467","last_reissued_at":"2026-06-10T00:08:30.989806Z","signature_status":"signed_v1","first_computed_at":"2026-06-10T00:08:30.989806Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Operator Fusion for LLM Inference on the Tensix Architecture","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jie Yu, Ke Li, Lili Liu, Qingbo Wu, Ruian Zhang, Wenzhu Wang","submitted_at":"2026-06-03T02:48:34Z","abstract_excerpt":"This study addresses on-device inference bottlenecks of Transformer models on Tenstorrent's Tensix architecture and proposes an operator fusion strategy that enhances data locality. RMSNorm is fused with matrix multiplication in self-attention and in the FFN, enabling back-to-back execution of memory-bound and compute-bound operators in on-chip SRAM to significantly reduce DRAM reads/writes of intermediate results and scheduling overhead. To support multi-core parallelism, a NoC-based multicast mechanism is leveraged in which row/column master nodes efficiently distribute inputs and weights ac"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.09879","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/2606.09879/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":"2606.09879","created_at":"2026-06-10T00:08:30.989972+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.09879v1","created_at":"2026-06-10T00:08:30.989972+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.09879","created_at":"2026-06-10T00:08:30.989972+00:00"},{"alias_kind":"pith_short_12","alias_value":"KN7ZSGHKBRNF","created_at":"2026-06-10T00:08:30.989972+00:00"},{"alias_kind":"pith_short_16","alias_value":"KN7ZSGHKBRNFE3TR","created_at":"2026-06-10T00:08:30.989972+00:00"},{"alias_kind":"pith_short_8","alias_value":"KN7ZSGHK","created_at":"2026-06-10T00:08:30.989972+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/KN7ZSGHKBRNFE3TREMZC4WNIRZ","json":"https://pith.science/pith/KN7ZSGHKBRNFE3TREMZC4WNIRZ.json","graph_json":"https://pith.science/api/pith-number/KN7ZSGHKBRNFE3TREMZC4WNIRZ/graph.json","events_json":"https://pith.science/api/pith-number/KN7ZSGHKBRNFE3TREMZC4WNIRZ/events.json","paper":"https://pith.science/paper/KN7ZSGHK"},"agent_actions":{"view_html":"https://pith.science/pith/KN7ZSGHKBRNFE3TREMZC4WNIRZ","download_json":"https://pith.science/pith/KN7ZSGHKBRNFE3TREMZC4WNIRZ.json","view_paper":"https://pith.science/paper/KN7ZSGHK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.09879&json=true","fetch_graph":"https://pith.science/api/pith-number/KN7ZSGHKBRNFE3TREMZC4WNIRZ/graph.json","fetch_events":"https://pith.science/api/pith-number/KN7ZSGHKBRNFE3TREMZC4WNIRZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KN7ZSGHKBRNFE3TREMZC4WNIRZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KN7ZSGHKBRNFE3TREMZC4WNIRZ/action/storage_attestation","attest_author":"https://pith.science/pith/KN7ZSGHKBRNFE3TREMZC4WNIRZ/action/author_attestation","sign_citation":"https://pith.science/pith/KN7ZSGHKBRNFE3TREMZC4WNIRZ/action/citation_signature","submit_replication":"https://pith.science/pith/KN7ZSGHKBRNFE3TREMZC4WNIRZ/action/replication_record"}},"created_at":"2026-06-10T00:08:30.989972+00:00","updated_at":"2026-06-10T00:08:30.989972+00:00"}