{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:ZZDC4WM2OJ5CREJ7PXMULCOD7N","short_pith_number":"pith:ZZDC4WM2","canonical_record":{"source":{"id":"2401.08671","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.PF","submitted_at":"2024-01-09T06:49:40Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"347faf36cb17aa9af70084b1d6cca28356920dfe4da0252edfa1b9ad0d39d597","abstract_canon_sha256":"14e7d91f6d462597c307ed42930ee697f646ec8270161329b646f6908dc7e777"},"schema_version":"1.0"},"canonical_sha256":"ce462e599a727a28913f7dd94589c3fb7f61f0106a94358dfa5a1fd9409a6fef","source":{"kind":"arxiv","id":"2401.08671","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.08671","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"arxiv_version","alias_value":"2401.08671v1","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.08671","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"pith_short_12","alias_value":"ZZDC4WM2OJ5C","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"pith_short_16","alias_value":"ZZDC4WM2OJ5CREJ7","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"pith_short_8","alias_value":"ZZDC4WM2","created_at":"2026-07-05T07:34:19Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:ZZDC4WM2OJ5CREJ7PXMULCOD7N","target":"record","payload":{"canonical_record":{"source":{"id":"2401.08671","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.PF","submitted_at":"2024-01-09T06:49:40Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"347faf36cb17aa9af70084b1d6cca28356920dfe4da0252edfa1b9ad0d39d597","abstract_canon_sha256":"14e7d91f6d462597c307ed42930ee697f646ec8270161329b646f6908dc7e777"},"schema_version":"1.0"},"canonical_sha256":"ce462e599a727a28913f7dd94589c3fb7f61f0106a94358dfa5a1fd9409a6fef","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:34:19.790390Z","signature_b64":"iTd/HP9gUGdO9ZFgNcEW6ujpQgWNLrs8j05vpv9AOJ8QtCAblHbnzMfSuyADBjZS+GNhmgdPAoC3WFZp/JWvDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ce462e599a727a28913f7dd94589c3fb7f61f0106a94358dfa5a1fd9409a6fef","last_reissued_at":"2026-07-05T07:34:19.790002Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:34:19.790002Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2401.08671","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:34:19Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SqGXbJlyLtlUCs2AGbz9oGMTGLDrVCVm7YSPty7RhZKriXgA+mZcxb91/5dEQWnV9Q+g54oeUj6TWlKbGUzmDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T15:37:35.016365Z"},"content_sha256":"0ec55e50207571b72638dc7e4b0aebc21f662b7fdc25890e78e20533287dc76b","schema_version":"1.0","event_id":"sha256:0ec55e50207571b72638dc7e4b0aebc21f662b7fdc25890e78e20533287dc76b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:ZZDC4WM2OJ5CREJ7PXMULCOD7N","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"DeepSpeed-FastGen: High-throughput Text Generation for LLMs via MII and DeepSpeed-Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.PF","authors_text":"Ammar Ahmad Awan, Arash Bakhtiari, Connor Holmes, Heyang Qin, Jeff Rasley, Lev Kurilenko, Masahiro Tanaka, Michael Wyatt, Reza Yazdani Aminabadi, Samyam Rajbhandari, Yuxiong He","submitted_at":"2024-01-09T06:49:40Z","abstract_excerpt":"The deployment and scaling of large language models (LLMs) have become critical as they permeate various applications, demanding high-throughput and low-latency serving systems. Existing frameworks struggle to balance these requirements, especially for workloads with long prompts. This paper introduces DeepSpeed-FastGen, a system that employs Dynamic SplitFuse, a novel prompt and generation composition strategy, to deliver up to 2.3x higher effective throughput, 2x lower latency on average, and up to 3.7x lower (token-level) tail latency, compared to state-of-the-art systems like vLLM. We leve"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.08671","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/2401.08671/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:34:19Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iGnN3Xh1gWB9zfb/ZRKtnBBnhlsEP/gf9tE+ruD4JJbgZWmbKE5K021aUw2a4SwpUQeqTbi5jueC+h2uW4L+Ag==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T15:37:35.017103Z"},"content_sha256":"2b8096231c74c99ebf700f3b836636c2361638e3ed4094e208c99c4fc6404f72","schema_version":"1.0","event_id":"sha256:2b8096231c74c99ebf700f3b836636c2361638e3ed4094e208c99c4fc6404f72"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ZZDC4WM2OJ5CREJ7PXMULCOD7N/bundle.json","state_url":"https://pith.science/pith/ZZDC4WM2OJ5CREJ7PXMULCOD7N/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ZZDC4WM2OJ5CREJ7PXMULCOD7N/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-04T15:37:35Z","links":{"resolver":"https://pith.science/pith/ZZDC4WM2OJ5CREJ7PXMULCOD7N","bundle":"https://pith.science/pith/ZZDC4WM2OJ5CREJ7PXMULCOD7N/bundle.json","state":"https://pith.science/pith/ZZDC4WM2OJ5CREJ7PXMULCOD7N/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ZZDC4WM2OJ5CREJ7PXMULCOD7N/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:ZZDC4WM2OJ5CREJ7PXMULCOD7N","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":"14e7d91f6d462597c307ed42930ee697f646ec8270161329b646f6908dc7e777","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.PF","submitted_at":"2024-01-09T06:49:40Z","title_canon_sha256":"347faf36cb17aa9af70084b1d6cca28356920dfe4da0252edfa1b9ad0d39d597"},"schema_version":"1.0","source":{"id":"2401.08671","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.08671","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"arxiv_version","alias_value":"2401.08671v1","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.08671","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"pith_short_12","alias_value":"ZZDC4WM2OJ5C","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"pith_short_16","alias_value":"ZZDC4WM2OJ5CREJ7","created_at":"2026-07-05T07:34:19Z"},{"alias_kind":"pith_short_8","alias_value":"ZZDC4WM2","created_at":"2026-07-05T07:34:19Z"}],"graph_snapshots":[{"event_id":"sha256:2b8096231c74c99ebf700f3b836636c2361638e3ed4094e208c99c4fc6404f72","target":"graph","created_at":"2026-07-05T07:34:19Z","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/2401.08671/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The deployment and scaling of large language models (LLMs) have become critical as they permeate various applications, demanding high-throughput and low-latency serving systems. Existing frameworks struggle to balance these requirements, especially for workloads with long prompts. This paper introduces DeepSpeed-FastGen, a system that employs Dynamic SplitFuse, a novel prompt and generation composition strategy, to deliver up to 2.3x higher effective throughput, 2x lower latency on average, and up to 3.7x lower (token-level) tail latency, compared to state-of-the-art systems like vLLM. We leve","authors_text":"Ammar Ahmad Awan, Arash Bakhtiari, Connor Holmes, Heyang Qin, Jeff Rasley, Lev Kurilenko, Masahiro Tanaka, Michael Wyatt, Reza Yazdani Aminabadi, Samyam Rajbhandari, Yuxiong He","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.PF","submitted_at":"2024-01-09T06:49:40Z","title":"DeepSpeed-FastGen: High-throughput Text Generation for LLMs via MII and DeepSpeed-Inference"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.08671","kind":"arxiv","version":1},"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":"sha256:0ec55e50207571b72638dc7e4b0aebc21f662b7fdc25890e78e20533287dc76b","target":"record","created_at":"2026-07-05T07:34:19Z","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":"14e7d91f6d462597c307ed42930ee697f646ec8270161329b646f6908dc7e777","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.PF","submitted_at":"2024-01-09T06:49:40Z","title_canon_sha256":"347faf36cb17aa9af70084b1d6cca28356920dfe4da0252edfa1b9ad0d39d597"},"schema_version":"1.0","source":{"id":"2401.08671","kind":"arxiv","version":1}},"canonical_sha256":"ce462e599a727a28913f7dd94589c3fb7f61f0106a94358dfa5a1fd9409a6fef","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ce462e599a727a28913f7dd94589c3fb7f61f0106a94358dfa5a1fd9409a6fef","first_computed_at":"2026-07-05T07:34:19.790002Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:34:19.790002Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"iTd/HP9gUGdO9ZFgNcEW6ujpQgWNLrs8j05vpv9AOJ8QtCAblHbnzMfSuyADBjZS+GNhmgdPAoC3WFZp/JWvDA==","signature_status":"signed_v1","signed_at":"2026-07-05T07:34:19.790390Z","signed_message":"canonical_sha256_bytes"},"source_id":"2401.08671","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0ec55e50207571b72638dc7e4b0aebc21f662b7fdc25890e78e20533287dc76b","sha256:2b8096231c74c99ebf700f3b836636c2361638e3ed4094e208c99c4fc6404f72"],"state_sha256":"9a9667e7fa7b45c125e6ffb6c1b015f3cef3630481fee740d8f4aac016342c62"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YM3UFIfD/w0tLSV5msnnKVwmRd52ySjuiAeslSwtUTBNcvWYo9yEwGjtD9TyRhg6A7YHav7VMxALR5YWZ4D6DA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T15:37:35.026120Z","bundle_sha256":"2a49eb3c780b1c8cf705a397c57c2fd2c42ece5611ecb1bc15c8ef4e294db1b2"}}