{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:IUWLQCSGFUWDECUSQPTIYEQ6ZQ","short_pith_number":"pith:IUWLQCSG","canonical_record":{"source":{"id":"2511.04791","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-11-06T20:18:34Z","cross_cats_sorted":[],"title_canon_sha256":"08a017afb2a6c36ac126bb9975114dc2b9d20b4cd0f7795d48dcf78ffa08a6ca","abstract_canon_sha256":"64ccf35c7aba444f8cbfad5c68d0d28474036eceec2be88b3b13303bd2170317"},"schema_version":"1.0"},"canonical_sha256":"452cb80a462d2c320a9283e68c121ecc1c175502e5107e0b812b9dc7ddd880de","source":{"kind":"arxiv","id":"2511.04791","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2511.04791","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"arxiv_version","alias_value":"2511.04791v2","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2511.04791","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"pith_short_12","alias_value":"IUWLQCSGFUWD","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"pith_short_16","alias_value":"IUWLQCSGFUWDECUS","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"pith_short_8","alias_value":"IUWLQCSG","created_at":"2026-06-02T01:03:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:IUWLQCSGFUWDECUSQPTIYEQ6ZQ","target":"record","payload":{"canonical_record":{"source":{"id":"2511.04791","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-11-06T20:18:34Z","cross_cats_sorted":[],"title_canon_sha256":"08a017afb2a6c36ac126bb9975114dc2b9d20b4cd0f7795d48dcf78ffa08a6ca","abstract_canon_sha256":"64ccf35c7aba444f8cbfad5c68d0d28474036eceec2be88b3b13303bd2170317"},"schema_version":"1.0"},"canonical_sha256":"452cb80a462d2c320a9283e68c121ecc1c175502e5107e0b812b9dc7ddd880de","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:37.951702Z","signature_b64":"RDRkzNrIAcsw1KW7fHShBhh+3e1IE9W0u4R/tUisyDogXYOYp2PZkjbMGqdQ1wsRjm3lVvWrwmqruSj0ngRfBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"452cb80a462d2c320a9283e68c121ecc1c175502e5107e0b812b9dc7ddd880de","last_reissued_at":"2026-06-02T01:03:37.951027Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:37.951027Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2511.04791","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-02T01:03:37Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9NE6Dyg7QXfZYoX+SK/UXKlsJkxPEFmtPjuKIaEXuVKUj3Nh62UhPhvGsad+qprVSTbspfnD75UBkgyiOlmpDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T12:46:01.556187Z"},"content_sha256":"d4f119b0e6eea485938c83833e4c050da59771e927a15ba6ce1f765943aa8710","schema_version":"1.0","event_id":"sha256:d4f119b0e6eea485938c83833e4c050da59771e927a15ba6ce1f765943aa8710"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:IUWLQCSGFUWDECUSQPTIYEQ6ZQ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"DuetServe: Harmonizing Prefill and Decode for LLM Serving via Adaptive GPU Multiplexing","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chaoyi Jiang, Daniel Wong, Hossein Entezari Zarch, Lei Gao, Mark Hill, Murali Annavaram","submitted_at":"2025-11-06T20:18:34Z","abstract_excerpt":"Modern LLM serving systems must sustain high throughput while meeting strict latency SLOs across two distinct inference phases: compute-intensive prefill and memory-bound decode phases. Existing approaches either (1) aggregate both phases on shared GPUs, leading to interference between prefill and decode phases, which degrades Time-Between-Tokens (TBT); or (2) disaggregate the two phases across GPUs, improving latency but wasting resources through duplicated models and KV cache transfers. We present DuetServe, a unified LLM serving framework that achieves disaggregation-level isolation within "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2511.04791","kind":"arxiv","version":2},"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/2511.04791/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":"2026-06-02T01:03:37Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ivmNvN1z58ihbRz1YnUCujR5tMrAC2MwaXmHN4pwoNZusWZ0rbRytGniK2Moz+bIlUybXvox2qA2TK/UrBhjCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T12:46:01.556567Z"},"content_sha256":"b75fa96c0204ce3c904422e72c58c01cb24fbec6e5e85e429b4887c57088e04a","schema_version":"1.0","event_id":"sha256:b75fa96c0204ce3c904422e72c58c01cb24fbec6e5e85e429b4887c57088e04a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IUWLQCSGFUWDECUSQPTIYEQ6ZQ/bundle.json","state_url":"https://pith.science/pith/IUWLQCSGFUWDECUSQPTIYEQ6ZQ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IUWLQCSGFUWDECUSQPTIYEQ6ZQ/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":"2026-06-04T12:46:01Z","links":{"resolver":"https://pith.science/pith/IUWLQCSGFUWDECUSQPTIYEQ6ZQ","bundle":"https://pith.science/pith/IUWLQCSGFUWDECUSQPTIYEQ6ZQ/bundle.json","state":"https://pith.science/pith/IUWLQCSGFUWDECUSQPTIYEQ6ZQ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IUWLQCSGFUWDECUSQPTIYEQ6ZQ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:IUWLQCSGFUWDECUSQPTIYEQ6ZQ","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":"64ccf35c7aba444f8cbfad5c68d0d28474036eceec2be88b3b13303bd2170317","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-11-06T20:18:34Z","title_canon_sha256":"08a017afb2a6c36ac126bb9975114dc2b9d20b4cd0f7795d48dcf78ffa08a6ca"},"schema_version":"1.0","source":{"id":"2511.04791","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2511.04791","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"arxiv_version","alias_value":"2511.04791v2","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2511.04791","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"pith_short_12","alias_value":"IUWLQCSGFUWD","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"pith_short_16","alias_value":"IUWLQCSGFUWDECUS","created_at":"2026-06-02T01:03:37Z"},{"alias_kind":"pith_short_8","alias_value":"IUWLQCSG","created_at":"2026-06-02T01:03:37Z"}],"graph_snapshots":[{"event_id":"sha256:b75fa96c0204ce3c904422e72c58c01cb24fbec6e5e85e429b4887c57088e04a","target":"graph","created_at":"2026-06-02T01:03:37Z","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/2511.04791/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Modern LLM serving systems must sustain high throughput while meeting strict latency SLOs across two distinct inference phases: compute-intensive prefill and memory-bound decode phases. Existing approaches either (1) aggregate both phases on shared GPUs, leading to interference between prefill and decode phases, which degrades Time-Between-Tokens (TBT); or (2) disaggregate the two phases across GPUs, improving latency but wasting resources through duplicated models and KV cache transfers. We present DuetServe, a unified LLM serving framework that achieves disaggregation-level isolation within ","authors_text":"Chaoyi Jiang, Daniel Wong, Hossein Entezari Zarch, Lei Gao, Mark Hill, Murali Annavaram","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-11-06T20:18:34Z","title":"DuetServe: Harmonizing Prefill and Decode for LLM Serving via Adaptive GPU Multiplexing"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2511.04791","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:d4f119b0e6eea485938c83833e4c050da59771e927a15ba6ce1f765943aa8710","target":"record","created_at":"2026-06-02T01:03:37Z","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":"64ccf35c7aba444f8cbfad5c68d0d28474036eceec2be88b3b13303bd2170317","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-11-06T20:18:34Z","title_canon_sha256":"08a017afb2a6c36ac126bb9975114dc2b9d20b4cd0f7795d48dcf78ffa08a6ca"},"schema_version":"1.0","source":{"id":"2511.04791","kind":"arxiv","version":2}},"canonical_sha256":"452cb80a462d2c320a9283e68c121ecc1c175502e5107e0b812b9dc7ddd880de","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"452cb80a462d2c320a9283e68c121ecc1c175502e5107e0b812b9dc7ddd880de","first_computed_at":"2026-06-02T01:03:37.951027Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T01:03:37.951027Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"RDRkzNrIAcsw1KW7fHShBhh+3e1IE9W0u4R/tUisyDogXYOYp2PZkjbMGqdQ1wsRjm3lVvWrwmqruSj0ngRfBA==","signature_status":"signed_v1","signed_at":"2026-06-02T01:03:37.951702Z","signed_message":"canonical_sha256_bytes"},"source_id":"2511.04791","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d4f119b0e6eea485938c83833e4c050da59771e927a15ba6ce1f765943aa8710","sha256:b75fa96c0204ce3c904422e72c58c01cb24fbec6e5e85e429b4887c57088e04a"],"state_sha256":"53f9595ac86a141ce65df653c43447bfa43aa65e001dc69e6c0d16cc408743cf"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xzCn/Ol8wyWye/NHD4kjpS3SsWpg/YeMvgZUdoShEtKtbeqMYKNFNPYiWXjLhwkzHVAToGd4nGKzOrQXlthwBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-04T12:46:01.558617Z","bundle_sha256":"47543ea9832cb34bb1f24d4633b029ee3e7c2a211682ecebd4d281dc0b4242ed"}}