{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TBEHZTOX5CDDPOQEHEY277KJRE","short_pith_number":"pith:TBEHZTOX","schema_version":"1.0","canonical_sha256":"98487ccdd7e88637ba043931affd49892eb980b4c83a9c8b15e37a4af695bec4","source":{"kind":"arxiv","id":"2508.16584","version":1},"attestation_state":"computed","paper":{"title":"TMA-Adaptive FP8 Grouped GEMM: Eliminating Padding Requirements in Low-Precision Training and Inference on Hopper","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Hui Wang, Jianfei Gao, Minxi Jin, Rong Fu, Weihan Cao, Zhilin Pei, Zhongling Su","submitted_at":"2025-08-07T03:24:31Z","abstract_excerpt":"Current FP8 grouped GEMM implementations require padding each group to a fixed alignment (e.g., 128), incurring memory and computational overhead. We propose \\textit{TMA-Adaptive FP8 Grouped GEMM}, which eliminates padding by dynamically adapting to variable group dimensions via (1) a TMA descriptor pool with $\\log_2(block_M)$ preconfigured descriptors to handle all residual row cases through dynamic runtime selection and dual-phase load-store operations, achieving comprehensive coverage with minimal overhead, and (2) TMA-alignment-aware management to satisfy 16-byte global memory alignment an"},"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":"2508.16584","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2025-08-07T03:24:31Z","cross_cats_sorted":[],"title_canon_sha256":"7d3ef81c47c203456d00043883c1766c63805da7886b69acdfafb0bb531af00f","abstract_canon_sha256":"17b3d5a609d637b76ce112f2d8911f8ccc63bdc6759e43b8229db1a4ca7dae88"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:57:59.678874Z","signature_b64":"LLKlKke6bmfnfotoI9D4V5gmzrv9eWblpdg/fgllpkQ6w1qNOHUhTaL47rd9chyKJ9UTbMN9sM9awgV4dg2eBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"98487ccdd7e88637ba043931affd49892eb980b4c83a9c8b15e37a4af695bec4","last_reissued_at":"2026-07-05T11:57:59.678431Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:57:59.678431Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TMA-Adaptive FP8 Grouped GEMM: Eliminating Padding Requirements in Low-Precision Training and Inference on Hopper","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Hui Wang, Jianfei Gao, Minxi Jin, Rong Fu, Weihan Cao, Zhilin Pei, Zhongling Su","submitted_at":"2025-08-07T03:24:31Z","abstract_excerpt":"Current FP8 grouped GEMM implementations require padding each group to a fixed alignment (e.g., 128), incurring memory and computational overhead. We propose \\textit{TMA-Adaptive FP8 Grouped GEMM}, which eliminates padding by dynamically adapting to variable group dimensions via (1) a TMA descriptor pool with $\\log_2(block_M)$ preconfigured descriptors to handle all residual row cases through dynamic runtime selection and dual-phase load-store operations, achieving comprehensive coverage with minimal overhead, and (2) TMA-alignment-aware management to satisfy 16-byte global memory alignment an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.16584","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/2508.16584/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":"2508.16584","created_at":"2026-07-05T11:57:59.678487+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.16584v1","created_at":"2026-07-05T11:57:59.678487+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.16584","created_at":"2026-07-05T11:57:59.678487+00:00"},{"alias_kind":"pith_short_12","alias_value":"TBEHZTOX5CDD","created_at":"2026-07-05T11:57:59.678487+00:00"},{"alias_kind":"pith_short_16","alias_value":"TBEHZTOX5CDDPOQE","created_at":"2026-07-05T11:57:59.678487+00:00"},{"alias_kind":"pith_short_8","alias_value":"TBEHZTOX","created_at":"2026-07-05T11:57:59.678487+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/TBEHZTOX5CDDPOQEHEY277KJRE","json":"https://pith.science/pith/TBEHZTOX5CDDPOQEHEY277KJRE.json","graph_json":"https://pith.science/api/pith-number/TBEHZTOX5CDDPOQEHEY277KJRE/graph.json","events_json":"https://pith.science/api/pith-number/TBEHZTOX5CDDPOQEHEY277KJRE/events.json","paper":"https://pith.science/paper/TBEHZTOX"},"agent_actions":{"view_html":"https://pith.science/pith/TBEHZTOX5CDDPOQEHEY277KJRE","download_json":"https://pith.science/pith/TBEHZTOX5CDDPOQEHEY277KJRE.json","view_paper":"https://pith.science/paper/TBEHZTOX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.16584&json=true","fetch_graph":"https://pith.science/api/pith-number/TBEHZTOX5CDDPOQEHEY277KJRE/graph.json","fetch_events":"https://pith.science/api/pith-number/TBEHZTOX5CDDPOQEHEY277KJRE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TBEHZTOX5CDDPOQEHEY277KJRE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TBEHZTOX5CDDPOQEHEY277KJRE/action/storage_attestation","attest_author":"https://pith.science/pith/TBEHZTOX5CDDPOQEHEY277KJRE/action/author_attestation","sign_citation":"https://pith.science/pith/TBEHZTOX5CDDPOQEHEY277KJRE/action/citation_signature","submit_replication":"https://pith.science/pith/TBEHZTOX5CDDPOQEHEY277KJRE/action/replication_record"}},"created_at":"2026-07-05T11:57:59.678487+00:00","updated_at":"2026-07-05T11:57:59.678487+00:00"}