{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:LIF6HK3OW4TUIBN4YR6TBWXFDZ","short_pith_number":"pith:LIF6HK3O","schema_version":"1.0","canonical_sha256":"5a0be3ab6eb7274405bcc47d30dae51e749b46e4f649e293e68bf52bf5486d63","source":{"kind":"arxiv","id":"2506.22169","version":1},"attestation_state":"computed","paper":{"title":"MCFuser: High-Performance and Rapid Fusion of Memory-Bound Compute-Intensive Operators","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.PL"],"primary_cat":"cs.DC","authors_text":"Dazhao Cheng, Donglin Yang, Xiaobo Zhou, Zheng Zhang","submitted_at":"2025-06-27T12:31:24Z","abstract_excerpt":"Operator fusion, a key technique to improve data locality and alleviate GPU memory bandwidth pressure, often fails to extend to the fusion of multiple compute-intensive operators due to saturated computation throughput. However, the dynamicity of tensor dimension sizes could potentially lead to these operators becoming memory-bound, necessitating the generation of fused kernels, a task hindered by limited search spaces for fusion strategies, redundant memory access, and prolonged tuning time, leading to sub-optimal performance and inefficient deployment.\n  We introduce MCFuser, a pioneering fr"},"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":"2506.22169","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.DC","submitted_at":"2025-06-27T12:31:24Z","cross_cats_sorted":["cs.PL"],"title_canon_sha256":"411f65fcddcd0583e60548952922e71edb5a136195e34f1fb694a22c579746dc","abstract_canon_sha256":"4771a021944f410cd6af8506820992485e9de2103f91087f027d1380b1655d8b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:28:23.050333Z","signature_b64":"YShBvS5pichjdvWbmw0aacR2Y9hhmz9EYZVx3K/652hd3KjGsoyLL9gJ0k3/n/jxRmKYB7baJUeEWlgXR04OAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5a0be3ab6eb7274405bcc47d30dae51e749b46e4f649e293e68bf52bf5486d63","last_reissued_at":"2026-07-05T11:28:23.049860Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:28:23.049860Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MCFuser: High-Performance and Rapid Fusion of Memory-Bound Compute-Intensive Operators","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.PL"],"primary_cat":"cs.DC","authors_text":"Dazhao Cheng, Donglin Yang, Xiaobo Zhou, Zheng Zhang","submitted_at":"2025-06-27T12:31:24Z","abstract_excerpt":"Operator fusion, a key technique to improve data locality and alleviate GPU memory bandwidth pressure, often fails to extend to the fusion of multiple compute-intensive operators due to saturated computation throughput. However, the dynamicity of tensor dimension sizes could potentially lead to these operators becoming memory-bound, necessitating the generation of fused kernels, a task hindered by limited search spaces for fusion strategies, redundant memory access, and prolonged tuning time, leading to sub-optimal performance and inefficient deployment.\n  We introduce MCFuser, a pioneering fr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.22169","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/2506.22169/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":"2506.22169","created_at":"2026-07-05T11:28:23.049914+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.22169v1","created_at":"2026-07-05T11:28:23.049914+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.22169","created_at":"2026-07-05T11:28:23.049914+00:00"},{"alias_kind":"pith_short_12","alias_value":"LIF6HK3OW4TU","created_at":"2026-07-05T11:28:23.049914+00:00"},{"alias_kind":"pith_short_16","alias_value":"LIF6HK3OW4TUIBN4","created_at":"2026-07-05T11:28:23.049914+00:00"},{"alias_kind":"pith_short_8","alias_value":"LIF6HK3O","created_at":"2026-07-05T11:28:23.049914+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/LIF6HK3OW4TUIBN4YR6TBWXFDZ","json":"https://pith.science/pith/LIF6HK3OW4TUIBN4YR6TBWXFDZ.json","graph_json":"https://pith.science/api/pith-number/LIF6HK3OW4TUIBN4YR6TBWXFDZ/graph.json","events_json":"https://pith.science/api/pith-number/LIF6HK3OW4TUIBN4YR6TBWXFDZ/events.json","paper":"https://pith.science/paper/LIF6HK3O"},"agent_actions":{"view_html":"https://pith.science/pith/LIF6HK3OW4TUIBN4YR6TBWXFDZ","download_json":"https://pith.science/pith/LIF6HK3OW4TUIBN4YR6TBWXFDZ.json","view_paper":"https://pith.science/paper/LIF6HK3O","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.22169&json=true","fetch_graph":"https://pith.science/api/pith-number/LIF6HK3OW4TUIBN4YR6TBWXFDZ/graph.json","fetch_events":"https://pith.science/api/pith-number/LIF6HK3OW4TUIBN4YR6TBWXFDZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LIF6HK3OW4TUIBN4YR6TBWXFDZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LIF6HK3OW4TUIBN4YR6TBWXFDZ/action/storage_attestation","attest_author":"https://pith.science/pith/LIF6HK3OW4TUIBN4YR6TBWXFDZ/action/author_attestation","sign_citation":"https://pith.science/pith/LIF6HK3OW4TUIBN4YR6TBWXFDZ/action/citation_signature","submit_replication":"https://pith.science/pith/LIF6HK3OW4TUIBN4YR6TBWXFDZ/action/replication_record"}},"created_at":"2026-07-05T11:28:23.049914+00:00","updated_at":"2026-07-05T11:28:23.049914+00:00"}