{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2FQL2UO6GZEE4JONVQBLJJWQ5D","short_pith_number":"pith:2FQL2UO6","schema_version":"1.0","canonical_sha256":"d160bd51de36484e25cdac02b4a6d0e8f850bd2b14c30e40fe93a713aebc6866","source":{"kind":"arxiv","id":"2605.20799","version":1},"attestation_state":"computed","paper":{"title":"Instant GPU Efficiency Visibility at Fleet Scale","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Collin Neale, Connor Pedersen, Dong H. Ahn, Michel Migdal, Nik Konyuchenko","submitted_at":"2026-05-20T06:44:25Z","abstract_excerpt":"We present Overall FLOP Utilization (OFU), a hardware-level, precision-agnostic GPU efficiency metric for AI workloads on HPC systems, derived from two on-chip performance counters: Tensor Pipe Activity and SM clock frequency. OFU requires no application instrumentation and works across GPU generations and numeric precisions. We characterize five properties of the OFU approximation -- tile quantization, floating-point precision scaling, clock sampling noise, Tensor Core clock domains, and non-tensor undercounting -- through controlled GEMM experiments on H100 and GB200 across FP16, TF32, FP8, "},"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":"2605.20799","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2026-05-20T06:44:25Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"911f8c8c795567b6812b6a6b95131b03084d9dfb75bfdb4c986acded710970ee","abstract_canon_sha256":"cfe4d46abee15952b102ff6eb3843eac2e0323c6e62df6d2d87497ec708d948e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:04:55.133941Z","signature_b64":"PGF+r1xb1pRTqawSVCVFVKKpxP4yO9cGj+QWy2Bo/w9GugdXIsQfmB27QRQsJzVSJFDEQgffT7a6BPoykrKRBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d160bd51de36484e25cdac02b4a6d0e8f850bd2b14c30e40fe93a713aebc6866","last_reissued_at":"2026-05-21T01:04:55.132996Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:04:55.132996Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Instant GPU Efficiency Visibility at Fleet Scale","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Collin Neale, Connor Pedersen, Dong H. Ahn, Michel Migdal, Nik Konyuchenko","submitted_at":"2026-05-20T06:44:25Z","abstract_excerpt":"We present Overall FLOP Utilization (OFU), a hardware-level, precision-agnostic GPU efficiency metric for AI workloads on HPC systems, derived from two on-chip performance counters: Tensor Pipe Activity and SM clock frequency. OFU requires no application instrumentation and works across GPU generations and numeric precisions. We characterize five properties of the OFU approximation -- tile quantization, floating-point precision scaling, clock sampling noise, Tensor Core clock domains, and non-tensor undercounting -- through controlled GEMM experiments on H100 and GB200 across FP16, TF32, FP8, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20799","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/2605.20799/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":"2605.20799","created_at":"2026-05-21T01:04:55.133130+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20799v1","created_at":"2026-05-21T01:04:55.133130+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20799","created_at":"2026-05-21T01:04:55.133130+00:00"},{"alias_kind":"pith_short_12","alias_value":"2FQL2UO6GZEE","created_at":"2026-05-21T01:04:55.133130+00:00"},{"alias_kind":"pith_short_16","alias_value":"2FQL2UO6GZEE4JON","created_at":"2026-05-21T01:04:55.133130+00:00"},{"alias_kind":"pith_short_8","alias_value":"2FQL2UO6","created_at":"2026-05-21T01:04:55.133130+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/2FQL2UO6GZEE4JONVQBLJJWQ5D","json":"https://pith.science/pith/2FQL2UO6GZEE4JONVQBLJJWQ5D.json","graph_json":"https://pith.science/api/pith-number/2FQL2UO6GZEE4JONVQBLJJWQ5D/graph.json","events_json":"https://pith.science/api/pith-number/2FQL2UO6GZEE4JONVQBLJJWQ5D/events.json","paper":"https://pith.science/paper/2FQL2UO6"},"agent_actions":{"view_html":"https://pith.science/pith/2FQL2UO6GZEE4JONVQBLJJWQ5D","download_json":"https://pith.science/pith/2FQL2UO6GZEE4JONVQBLJJWQ5D.json","view_paper":"https://pith.science/paper/2FQL2UO6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20799&json=true","fetch_graph":"https://pith.science/api/pith-number/2FQL2UO6GZEE4JONVQBLJJWQ5D/graph.json","fetch_events":"https://pith.science/api/pith-number/2FQL2UO6GZEE4JONVQBLJJWQ5D/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2FQL2UO6GZEE4JONVQBLJJWQ5D/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2FQL2UO6GZEE4JONVQBLJJWQ5D/action/storage_attestation","attest_author":"https://pith.science/pith/2FQL2UO6GZEE4JONVQBLJJWQ5D/action/author_attestation","sign_citation":"https://pith.science/pith/2FQL2UO6GZEE4JONVQBLJJWQ5D/action/citation_signature","submit_replication":"https://pith.science/pith/2FQL2UO6GZEE4JONVQBLJJWQ5D/action/replication_record"}},"created_at":"2026-05-21T01:04:55.133130+00:00","updated_at":"2026-05-21T01:04:55.133130+00:00"}