{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:JNDKID67RZ2DDFMISHSELLTN7U","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"5a2b3c2621efdfee38b0d3fd93d1257f23448152c17c09a0640c0f16da189999","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2024-01-05T19:17:55Z","title_canon_sha256":"2f730c10376a3cbab5771346e0daf7dffb516d705214dd42ff89d23ffeb2b7db"},"schema_version":"1.0","source":{"id":"2402.00025","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.00025","created_at":"2026-07-05T07:48:17Z"},{"alias_kind":"arxiv_version","alias_value":"2402.00025v2","created_at":"2026-07-05T07:48:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.00025","created_at":"2026-07-05T07:48:17Z"},{"alias_kind":"pith_short_12","alias_value":"JNDKID67RZ2D","created_at":"2026-07-05T07:48:17Z"},{"alias_kind":"pith_short_16","alias_value":"JNDKID67RZ2DDFMI","created_at":"2026-07-05T07:48:17Z"},{"alias_kind":"pith_short_8","alias_value":"JNDKID67","created_at":"2026-07-05T07:48:17Z"}],"graph_snapshots":[{"event_id":"sha256:90e69f57c74cb2667ca983a6ef12d73f226b49c8d96456764a30b8db9685e4a4","target":"graph","created_at":"2026-07-05T07:48:17Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2402.00025/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We propose an implementation of an efficient fused matrix multiplication kernel for W4A16 quantized inference, where we perform dequantization and GEMM in a fused kernel using a SplitK work decomposition. Our implementation shows improvement for the type of skinny matrix-matrix multiplications found in foundation model inference workloads. In particular, this paper surveys the type of matrix multiplication between a skinny activation matrix and a square weight matrix. Our results show an average of 65% speed improvement on A100, and an average of 124% speed improvement on H100 (with a peak of ","authors_text":"Adnan Hoque, Chih-Chieh Yang, Less Wright, Mudhakar Srivatsa, Raghu Ganti","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2024-01-05T19:17:55Z","title":"Accelerating a Triton Fused Kernel for W4A16 Quantized Inference with SplitK work decomposition"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.00025","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:7843e63f7f3a0bd09449a1fbcdc48ae44dc75c2c26e76d4115ee6b66db54f790","target":"record","created_at":"2026-07-05T07:48:17Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"5a2b3c2621efdfee38b0d3fd93d1257f23448152c17c09a0640c0f16da189999","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2024-01-05T19:17:55Z","title_canon_sha256":"2f730c10376a3cbab5771346e0daf7dffb516d705214dd42ff89d23ffeb2b7db"},"schema_version":"1.0","source":{"id":"2402.00025","kind":"arxiv","version":2}},"canonical_sha256":"4b46a40fdf8e7431958891e445ae6dfd2232e4f1b18a4f4001c84d68818186b9","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4b46a40fdf8e7431958891e445ae6dfd2232e4f1b18a4f4001c84d68818186b9","first_computed_at":"2026-07-05T07:48:17.924202Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:48:17.924202Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Irkd4Ayz74j0DEpZVYrndRocQAv2w7ZoYfYHP4QFyIzQJvyglRuPAljQ885MjzyjZrYfTQ/F3rdoOQSs2UgoDg==","signature_status":"signed_v1","signed_at":"2026-07-05T07:48:17.924701Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.00025","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7843e63f7f3a0bd09449a1fbcdc48ae44dc75c2c26e76d4115ee6b66db54f790","sha256:90e69f57c74cb2667ca983a6ef12d73f226b49c8d96456764a30b8db9685e4a4"],"state_sha256":"70155c283cfa07996b441250a1890de1f64424e9c17562a7501067d277cb275f"}