{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:BMZWCNLMMQVEEGZJO7BC5FOIBQ","short_pith_number":"pith:BMZWCNLM","canonical_record":{"source":{"id":"2506.11309","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-06-12T21:15:58Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"4b6f32d95177a2cdb2c5512c072ae820ae844a5538c5e787aa864b08fe53c74f","abstract_canon_sha256":"f32a69ab6cb1d7b20f86020f8750cbb534ad99acf54f119c3bfa732eb0e1be3d"},"schema_version":"1.0"},"canonical_sha256":"0b3361356c642a421b2977c22e95c80c3906459f2972f736805d87fc3bb21d13","source":{"kind":"arxiv","id":"2506.11309","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.11309","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"arxiv_version","alias_value":"2506.11309v1","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.11309","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"pith_short_12","alias_value":"BMZWCNLMMQVE","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"pith_short_16","alias_value":"BMZWCNLMMQVEEGZJ","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"pith_short_8","alias_value":"BMZWCNLM","created_at":"2026-07-05T11:21:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:BMZWCNLMMQVEEGZJO7BC5FOIBQ","target":"record","payload":{"canonical_record":{"source":{"id":"2506.11309","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-06-12T21:15:58Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"4b6f32d95177a2cdb2c5512c072ae820ae844a5538c5e787aa864b08fe53c74f","abstract_canon_sha256":"f32a69ab6cb1d7b20f86020f8750cbb534ad99acf54f119c3bfa732eb0e1be3d"},"schema_version":"1.0"},"canonical_sha256":"0b3361356c642a421b2977c22e95c80c3906459f2972f736805d87fc3bb21d13","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:21:02.380610Z","signature_b64":"tg4OL4HNoZv57qoOIdOsGgS731S/sQz8tjLg2ds8Nc4Xc5kqCGeCG+LXks1fG6u4MHP0RqXMU3/g31lODtwQDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0b3361356c642a421b2977c22e95c80c3906459f2972f736805d87fc3bb21d13","last_reissued_at":"2026-07-05T11:21:02.380090Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:21:02.380090Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.11309","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":"2026-07-05T11:21:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HJgWx9NtFiYNzHYq2wJ7KiR4y4hMxvTCnkEd1xmOrAaBPX9XNqHFU8J3oOO9bHx0YmiPsbZ3inYED2y1SPEJBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T17:03:18.293140Z"},"content_sha256":"88e2d558782202fd082519f20bae0e426c80486e739b3762bde669cabb1021f2","schema_version":"1.0","event_id":"sha256:88e2d558782202fd082519f20bae0e426c80486e739b3762bde669cabb1021f2"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:BMZWCNLMMQVEEGZJO7BC5FOIBQ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SwiftSpec: Ultra-Low Latency LLM Decoding by Scaling Asynchronous Speculative Decoding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Chengquan Jiang, Haibin Lin, Henry Hoffmann, Menghan Yu, Size Zheng, Xin Liu, Ziheng Jiang, Ziyi Zhang","submitted_at":"2025-06-12T21:15:58Z","abstract_excerpt":"Low-latency decoding for large language models (LLMs) is crucial for applications like chatbots and code assistants, yet generating long outputs remains slow in single-query settings. Prior work on speculative decoding (which combines a small draft model with a larger target model) and tensor parallelism has each accelerated decoding. However, conventional approaches fail to apply both simultaneously due to imbalanced compute requirements (between draft and target models), KV-cache inconsistencies, and communication overheads under small-batch tensor-parallelism. This paper introduces SwiftSpe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.11309","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/2506.11309/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":"2026-07-05T11:21:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JcU5H1yFosVhn+BhxscvqmCscFy9yZuhT/fDBuCPMDWZf0UknU4WWdGlUVZQU19oaNh+9LjdybKxI1g/TNBJBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T17:03:18.293635Z"},"content_sha256":"1ad2c2f1110f5ed65573d1987f669faf37763557e0ef44b8cf8a64e5633a1a0b","schema_version":"1.0","event_id":"sha256:1ad2c2f1110f5ed65573d1987f669faf37763557e0ef44b8cf8a64e5633a1a0b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BMZWCNLMMQVEEGZJO7BC5FOIBQ/bundle.json","state_url":"https://pith.science/pith/BMZWCNLMMQVEEGZJO7BC5FOIBQ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BMZWCNLMMQVEEGZJO7BC5FOIBQ/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":"2026-08-04T17:03:18Z","links":{"resolver":"https://pith.science/pith/BMZWCNLMMQVEEGZJO7BC5FOIBQ","bundle":"https://pith.science/pith/BMZWCNLMMQVEEGZJO7BC5FOIBQ/bundle.json","state":"https://pith.science/pith/BMZWCNLMMQVEEGZJO7BC5FOIBQ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BMZWCNLMMQVEEGZJO7BC5FOIBQ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:BMZWCNLMMQVEEGZJO7BC5FOIBQ","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":"f32a69ab6cb1d7b20f86020f8750cbb534ad99acf54f119c3bfa732eb0e1be3d","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-06-12T21:15:58Z","title_canon_sha256":"4b6f32d95177a2cdb2c5512c072ae820ae844a5538c5e787aa864b08fe53c74f"},"schema_version":"1.0","source":{"id":"2506.11309","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.11309","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"arxiv_version","alias_value":"2506.11309v1","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.11309","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"pith_short_12","alias_value":"BMZWCNLMMQVE","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"pith_short_16","alias_value":"BMZWCNLMMQVEEGZJ","created_at":"2026-07-05T11:21:02Z"},{"alias_kind":"pith_short_8","alias_value":"BMZWCNLM","created_at":"2026-07-05T11:21:02Z"}],"graph_snapshots":[{"event_id":"sha256:1ad2c2f1110f5ed65573d1987f669faf37763557e0ef44b8cf8a64e5633a1a0b","target":"graph","created_at":"2026-07-05T11:21:02Z","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/2506.11309/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Low-latency decoding for large language models (LLMs) is crucial for applications like chatbots and code assistants, yet generating long outputs remains slow in single-query settings. Prior work on speculative decoding (which combines a small draft model with a larger target model) and tensor parallelism has each accelerated decoding. However, conventional approaches fail to apply both simultaneously due to imbalanced compute requirements (between draft and target models), KV-cache inconsistencies, and communication overheads under small-batch tensor-parallelism. This paper introduces SwiftSpe","authors_text":"Chengquan Jiang, Haibin Lin, Henry Hoffmann, Menghan Yu, Size Zheng, Xin Liu, Ziheng Jiang, Ziyi Zhang","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-06-12T21:15:58Z","title":"SwiftSpec: Ultra-Low Latency LLM Decoding by Scaling Asynchronous Speculative Decoding"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.11309","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":"sha256:88e2d558782202fd082519f20bae0e426c80486e739b3762bde669cabb1021f2","target":"record","created_at":"2026-07-05T11:21:02Z","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":"f32a69ab6cb1d7b20f86020f8750cbb534ad99acf54f119c3bfa732eb0e1be3d","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-06-12T21:15:58Z","title_canon_sha256":"4b6f32d95177a2cdb2c5512c072ae820ae844a5538c5e787aa864b08fe53c74f"},"schema_version":"1.0","source":{"id":"2506.11309","kind":"arxiv","version":1}},"canonical_sha256":"0b3361356c642a421b2977c22e95c80c3906459f2972f736805d87fc3bb21d13","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0b3361356c642a421b2977c22e95c80c3906459f2972f736805d87fc3bb21d13","first_computed_at":"2026-07-05T11:21:02.380090Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:21:02.380090Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"tg4OL4HNoZv57qoOIdOsGgS731S/sQz8tjLg2ds8Nc4Xc5kqCGeCG+LXks1fG6u4MHP0RqXMU3/g31lODtwQDg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:21:02.380610Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.11309","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:88e2d558782202fd082519f20bae0e426c80486e739b3762bde669cabb1021f2","sha256:1ad2c2f1110f5ed65573d1987f669faf37763557e0ef44b8cf8a64e5633a1a0b"],"state_sha256":"61388bb72b3d609b1f2d4c027e65db528163998b4ba2834b06ce78bd2890fa17"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oqSNUbLX6nkGVRBuW83wcqglizGNy2uNoA+rkTQ42mUmoYchDRyGrQDWLDo8HKJ/Iszg/ReqBg1noajtzU5LDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T17:03:18.297354Z","bundle_sha256":"73bc13a02ac49a9d1ac96d9f934933e72a32eb9d201e21cdec685834dabe764c"}}