{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:2IAX2M2YAOXPWGYGGJ7HTWNENH","short_pith_number":"pith:2IAX2M2Y","canonical_record":{"source":{"id":"2508.15601","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-08-21T14:24:52Z","cross_cats_sorted":["cs.PF"],"title_canon_sha256":"fd99dfd641caad1553c5a2ee174022657fcbf25b44fd3db30d84aade1e20d960","abstract_canon_sha256":"4465555fe84f2caea1ca592ed5630d52b5ac8f41b12886434ea538c50e80f226"},"schema_version":"1.0"},"canonical_sha256":"d2017d335803aefb1b06327e79d9a469c63a3b73f3198514e11f059dc54d6f80","source":{"kind":"arxiv","id":"2508.15601","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.15601","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"arxiv_version","alias_value":"2508.15601v2","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.15601","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"pith_short_12","alias_value":"2IAX2M2YAOXP","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"pith_short_16","alias_value":"2IAX2M2YAOXPWGYG","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"pith_short_8","alias_value":"2IAX2M2Y","created_at":"2026-05-20T00:01:35Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:2IAX2M2YAOXPWGYGGJ7HTWNENH","target":"record","payload":{"canonical_record":{"source":{"id":"2508.15601","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-08-21T14:24:52Z","cross_cats_sorted":["cs.PF"],"title_canon_sha256":"fd99dfd641caad1553c5a2ee174022657fcbf25b44fd3db30d84aade1e20d960","abstract_canon_sha256":"4465555fe84f2caea1ca592ed5630d52b5ac8f41b12886434ea538c50e80f226"},"schema_version":"1.0"},"canonical_sha256":"d2017d335803aefb1b06327e79d9a469c63a3b73f3198514e11f059dc54d6f80","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:01:35.253760Z","signature_b64":"DGG3q6j+MhK5B/UATLPmWo2hyz1oIuD/XtmipTZX99R70QjDS9P2TVgqwpyblQt2DzM+fe+EFy5NCPWMMck4DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d2017d335803aefb1b06327e79d9a469c63a3b73f3198514e11f059dc54d6f80","last_reissued_at":"2026-05-20T00:01:35.252971Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:01:35.252971Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2508.15601","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":"2026-05-20T00:01:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WUa1R1BThNPYCnt4MlorThQ5AnhaNJ0XttyeXiaL8sBbFx+sLubb6k4P+3PHpskmunPODvMvPY3bBMv/b9kICg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-24T16:43:34.033231Z"},"content_sha256":"22edcce606aef3dd10701dc43a29abb51e6fe7e58a7edd2deaaee30304014701","schema_version":"1.0","event_id":"sha256:22edcce606aef3dd10701dc43a29abb51e6fe7e58a7edd2deaaee30304014701"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:2IAX2M2YAOXPWGYGGJ7HTWNENH","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"LMDeploy Accelerates Mixed-Precision LLM Inference with TurboMind","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.PF"],"primary_cat":"cs.DC","authors_text":"Fangcheng Fu, Guoliang He, Han Lv, Kai Chen, Li Zhang, Ningsheng Ma, Qian Yao, Xin Chen, Youhe Jiang","submitted_at":"2025-08-21T14:24:52Z","abstract_excerpt":"Mixed-precision inference techniques reduce the memory and computational demands of Large Language Models (LLMs) by applying hybrid precision formats to model weights, activations, and KV caches. However, existing systems struggle to (i) automatically generalize across diverse hardware architectures and precision formats, often requiring fragmented, hand-tuned kernels, and (ii) fully exploit available memory and compute resources, often causing performance bottlenecks. To address these problems, we propose TurboMind, a generalizable and efficient mixed-precision LLM inference engine of LMDeplo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.15601","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/2508.15601/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":"2026-05-20T00:01:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1fhP9BzwnXDDKPkzuHxqBvDXSDQvV1kYWZ1Ep6q+XCFCMLqlxP4NHjUx9rneKABtHxY0aBg3znNX4ShnrNnHDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-24T16:43:34.034009Z"},"content_sha256":"b5aecce952944e3b05550931ec09511843af0275e27d3e691c14e717c6d96460","schema_version":"1.0","event_id":"sha256:b5aecce952944e3b05550931ec09511843af0275e27d3e691c14e717c6d96460"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/2IAX2M2YAOXPWGYGGJ7HTWNENH/bundle.json","state_url":"https://pith.science/pith/2IAX2M2YAOXPWGYGGJ7HTWNENH/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/2IAX2M2YAOXPWGYGGJ7HTWNENH/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":"2026-05-24T16:43:34Z","links":{"resolver":"https://pith.science/pith/2IAX2M2YAOXPWGYGGJ7HTWNENH","bundle":"https://pith.science/pith/2IAX2M2YAOXPWGYGGJ7HTWNENH/bundle.json","state":"https://pith.science/pith/2IAX2M2YAOXPWGYGGJ7HTWNENH/state.json","well_known_bundle":"https://pith.science/.well-known/pith/2IAX2M2YAOXPWGYGGJ7HTWNENH/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:2IAX2M2YAOXPWGYGGJ7HTWNENH","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":"4465555fe84f2caea1ca592ed5630d52b5ac8f41b12886434ea538c50e80f226","cross_cats_sorted":["cs.PF"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-08-21T14:24:52Z","title_canon_sha256":"fd99dfd641caad1553c5a2ee174022657fcbf25b44fd3db30d84aade1e20d960"},"schema_version":"1.0","source":{"id":"2508.15601","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.15601","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"arxiv_version","alias_value":"2508.15601v2","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.15601","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"pith_short_12","alias_value":"2IAX2M2YAOXP","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"pith_short_16","alias_value":"2IAX2M2YAOXPWGYG","created_at":"2026-05-20T00:01:35Z"},{"alias_kind":"pith_short_8","alias_value":"2IAX2M2Y","created_at":"2026-05-20T00:01:35Z"}],"graph_snapshots":[{"event_id":"sha256:b5aecce952944e3b05550931ec09511843af0275e27d3e691c14e717c6d96460","target":"graph","created_at":"2026-05-20T00:01:35Z","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/2508.15601/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Mixed-precision inference techniques reduce the memory and computational demands of Large Language Models (LLMs) by applying hybrid precision formats to model weights, activations, and KV caches. However, existing systems struggle to (i) automatically generalize across diverse hardware architectures and precision formats, often requiring fragmented, hand-tuned kernels, and (ii) fully exploit available memory and compute resources, often causing performance bottlenecks. To address these problems, we propose TurboMind, a generalizable and efficient mixed-precision LLM inference engine of LMDeplo","authors_text":"Fangcheng Fu, Guoliang He, Han Lv, Kai Chen, Li Zhang, Ningsheng Ma, Qian Yao, Xin Chen, Youhe Jiang","cross_cats":["cs.PF"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-08-21T14:24:52Z","title":"LMDeploy Accelerates Mixed-Precision LLM Inference with TurboMind"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.15601","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":"sha256:22edcce606aef3dd10701dc43a29abb51e6fe7e58a7edd2deaaee30304014701","target":"record","created_at":"2026-05-20T00:01:35Z","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":"4465555fe84f2caea1ca592ed5630d52b5ac8f41b12886434ea538c50e80f226","cross_cats_sorted":["cs.PF"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2025-08-21T14:24:52Z","title_canon_sha256":"fd99dfd641caad1553c5a2ee174022657fcbf25b44fd3db30d84aade1e20d960"},"schema_version":"1.0","source":{"id":"2508.15601","kind":"arxiv","version":2}},"canonical_sha256":"d2017d335803aefb1b06327e79d9a469c63a3b73f3198514e11f059dc54d6f80","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d2017d335803aefb1b06327e79d9a469c63a3b73f3198514e11f059dc54d6f80","first_computed_at":"2026-05-20T00:01:35.252971Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:01:35.252971Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"DGG3q6j+MhK5B/UATLPmWo2hyz1oIuD/XtmipTZX99R70QjDS9P2TVgqwpyblQt2DzM+fe+EFy5NCPWMMck4DQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:01:35.253760Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.15601","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:22edcce606aef3dd10701dc43a29abb51e6fe7e58a7edd2deaaee30304014701","sha256:b5aecce952944e3b05550931ec09511843af0275e27d3e691c14e717c6d96460"],"state_sha256":"9c18bd1a5c274fc226d736a5161540683e6d2e1703bcfbc727c3e8577d75cdaa"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ITzaqyf44LN72XpkxROvMVRi8q8BbhpHStR0wcEjuQTrUa/uS7oKVEq688arV5406Q0+1H6BYVjw2Pnlj8hNAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-24T16:43:34.039118Z","bundle_sha256":"467f95d7c3fd4d5d694d505409a208e9391a49240c7cfb40d4bedc1d32621fc5"}}