{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:AO6X2YC4ZRATM34XH425RR3IU2","short_pith_number":"pith:AO6X2YC4","canonical_record":{"source":{"id":"2404.14527","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2024-04-22T18:56:18Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"ab36325166569bef72eab1cb336afea03b0a40df302d0ed314a243d6c86ff60e","abstract_canon_sha256":"ac92c1dc844b6d537646902c4f3e067516d3ae0b9a1ad457df2394b927449620"},"schema_version":"1.0"},"canonical_sha256":"03bd7d605ccc41366f973f35d8c768a68e21c646ab6febb901b6d45422bcca80","source":{"kind":"arxiv","id":"2404.14527","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.14527","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"arxiv_version","alias_value":"2404.14527v4","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.14527","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"pith_short_12","alias_value":"AO6X2YC4ZRAT","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"pith_short_16","alias_value":"AO6X2YC4ZRATM34X","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"pith_short_8","alias_value":"AO6X2YC4","created_at":"2026-07-05T08:46:54Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:AO6X2YC4ZRATM34XH425RR3IU2","target":"record","payload":{"canonical_record":{"source":{"id":"2404.14527","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2024-04-22T18:56:18Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"ab36325166569bef72eab1cb336afea03b0a40df302d0ed314a243d6c86ff60e","abstract_canon_sha256":"ac92c1dc844b6d537646902c4f3e067516d3ae0b9a1ad457df2394b927449620"},"schema_version":"1.0"},"canonical_sha256":"03bd7d605ccc41366f973f35d8c768a68e21c646ab6febb901b6d45422bcca80","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:46:54.677590Z","signature_b64":"7wJLBdPlFsQppiiDZhI5C2q3aXCfOcEz2XoZWuWb199zF2CdQp1h9vmKbioYVN8WH35w0o/XsnjaExE6YxAnCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"03bd7d605ccc41366f973f35d8c768a68e21c646ab6febb901b6d45422bcca80","last_reissued_at":"2026-07-05T08:46:54.676994Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:46:54.676994Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.14527","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:46:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wCPypU2lR2TrrhFQQAwm4TLgXE0xJKg237sJM9RiWvjV3Ka7PBh4RATrQaYDp+CxtsjG1/XcFA7+9kFvS2EfAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-01T08:30:54.626942Z"},"content_sha256":"3485de78172cf16e4827f28318c4ee75a67700d0f893976cbf2e8c3586ec3030","schema_version":"1.0","event_id":"sha256:3485de78172cf16e4827f28318c4ee75a67700d0f893976cbf2e8c3586ec3030"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:AO6X2YC4ZRATM34XH425RR3IU2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"M\\'elange: Cost Efficient Large Language Model Serving by Exploiting GPU Heterogeneity","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Alvin Cheung, Doyoung Kim, Ion Stoica, Jiaxiang Yu, Tyler Griggs, Wei-Lin Chiang, Xiaoxuan Liu","submitted_at":"2024-04-22T18:56:18Z","abstract_excerpt":"Large language models (LLMs) are increasingly integrated into many online services, yet they remain cost-prohibitive to deploy due to the requirement of expensive GPU instances. Prior work has addressed the high cost of LLM serving by improving the inference engine, but less attention has been given to selecting the most cost-efficient GPU type(s) for a specific LLM service. There is a large and growing landscape of GPU types and, within these options, higher cost does not always lead to increased performance. Instead, through a comprehensive investigation, we find that three key LLM service c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.14527","kind":"arxiv","version":4},"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/2404.14527/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":"2026-07-05T08:46:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0Wtf+8EErEETwB57PpVHAfl+xehPm5FvDmOfUHSDIfiTR0WrLXM95IcWP4HLTKY0IAFsA9y/eb5QU4MOetMGCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-01T08:30:54.627414Z"},"content_sha256":"94251eaccb15ea8541cab8c0c29087527158b4932bf2547ca06f667960a2d6f7","schema_version":"1.0","event_id":"sha256:94251eaccb15ea8541cab8c0c29087527158b4932bf2547ca06f667960a2d6f7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/AO6X2YC4ZRATM34XH425RR3IU2/bundle.json","state_url":"https://pith.science/pith/AO6X2YC4ZRATM34XH425RR3IU2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/AO6X2YC4ZRATM34XH425RR3IU2/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":"2026-08-01T08:30:54Z","links":{"resolver":"https://pith.science/pith/AO6X2YC4ZRATM34XH425RR3IU2","bundle":"https://pith.science/pith/AO6X2YC4ZRATM34XH425RR3IU2/bundle.json","state":"https://pith.science/pith/AO6X2YC4ZRATM34XH425RR3IU2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/AO6X2YC4ZRATM34XH425RR3IU2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:AO6X2YC4ZRATM34XH425RR3IU2","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":"ac92c1dc844b6d537646902c4f3e067516d3ae0b9a1ad457df2394b927449620","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2024-04-22T18:56:18Z","title_canon_sha256":"ab36325166569bef72eab1cb336afea03b0a40df302d0ed314a243d6c86ff60e"},"schema_version":"1.0","source":{"id":"2404.14527","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.14527","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"arxiv_version","alias_value":"2404.14527v4","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.14527","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"pith_short_12","alias_value":"AO6X2YC4ZRAT","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"pith_short_16","alias_value":"AO6X2YC4ZRATM34X","created_at":"2026-07-05T08:46:54Z"},{"alias_kind":"pith_short_8","alias_value":"AO6X2YC4","created_at":"2026-07-05T08:46:54Z"}],"graph_snapshots":[{"event_id":"sha256:94251eaccb15ea8541cab8c0c29087527158b4932bf2547ca06f667960a2d6f7","target":"graph","created_at":"2026-07-05T08:46:54Z","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/2404.14527/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) are increasingly integrated into many online services, yet they remain cost-prohibitive to deploy due to the requirement of expensive GPU instances. Prior work has addressed the high cost of LLM serving by improving the inference engine, but less attention has been given to selecting the most cost-efficient GPU type(s) for a specific LLM service. There is a large and growing landscape of GPU types and, within these options, higher cost does not always lead to increased performance. Instead, through a comprehensive investigation, we find that three key LLM service c","authors_text":"Alvin Cheung, Doyoung Kim, Ion Stoica, Jiaxiang Yu, Tyler Griggs, Wei-Lin Chiang, Xiaoxuan Liu","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2024-04-22T18:56:18Z","title":"M\\'elange: Cost Efficient Large Language Model Serving by Exploiting GPU Heterogeneity"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.14527","kind":"arxiv","version":4},"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":"sha256:3485de78172cf16e4827f28318c4ee75a67700d0f893976cbf2e8c3586ec3030","target":"record","created_at":"2026-07-05T08:46:54Z","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":"ac92c1dc844b6d537646902c4f3e067516d3ae0b9a1ad457df2394b927449620","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2024-04-22T18:56:18Z","title_canon_sha256":"ab36325166569bef72eab1cb336afea03b0a40df302d0ed314a243d6c86ff60e"},"schema_version":"1.0","source":{"id":"2404.14527","kind":"arxiv","version":4}},"canonical_sha256":"03bd7d605ccc41366f973f35d8c768a68e21c646ab6febb901b6d45422bcca80","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"03bd7d605ccc41366f973f35d8c768a68e21c646ab6febb901b6d45422bcca80","first_computed_at":"2026-07-05T08:46:54.676994Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:46:54.676994Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"7wJLBdPlFsQppiiDZhI5C2q3aXCfOcEz2XoZWuWb199zF2CdQp1h9vmKbioYVN8WH35w0o/XsnjaExE6YxAnCw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:46:54.677590Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.14527","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3485de78172cf16e4827f28318c4ee75a67700d0f893976cbf2e8c3586ec3030","sha256:94251eaccb15ea8541cab8c0c29087527158b4932bf2547ca06f667960a2d6f7"],"state_sha256":"6b8690b5143807e258f00546a9282372d313fd34b3e54d16bf652c1d954448b3"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"q6aojdyNonY0OCCjYm9DnZwOwV3MesyE0W5ccXKrqU69/MOKNvw7Jn2QA4/IBh5h4D3vabZcxg9dwPLTriSMCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-01T08:30:54.631282Z","bundle_sha256":"9f411e76cfab5d0a8fe4b780c57308304e4697d1fc390a1aaef60903d1f07bd1"}}