{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:LOHGU5AX7BHMWC5MXWWMKZIY2K","short_pith_number":"pith:LOHGU5AX","canonical_record":{"source":{"id":"2507.10392","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.DC","submitted_at":"2025-07-14T15:31:31Z","cross_cats_sorted":[],"title_canon_sha256":"b0d75d25cf32a498adb46b18b5b4c6050510265b604e1f24be56a328c2c9de88","abstract_canon_sha256":"1bfe8bbd3a73debcb38d2c1e752b45cad5d44e99250cf9a15442c36a402ec375"},"schema_version":"1.0"},"canonical_sha256":"5b8e6a7417f84ecb0bacbdacc56518d2a4deccea0d5c9e7a1e718519261c9961","source":{"kind":"arxiv","id":"2507.10392","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.10392","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"arxiv_version","alias_value":"2507.10392v1","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.10392","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"pith_short_12","alias_value":"LOHGU5AX7BHM","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"pith_short_16","alias_value":"LOHGU5AX7BHMWC5M","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"pith_short_8","alias_value":"LOHGU5AX","created_at":"2026-07-05T11:36:50Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:LOHGU5AX7BHMWC5MXWWMKZIY2K","target":"record","payload":{"canonical_record":{"source":{"id":"2507.10392","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.DC","submitted_at":"2025-07-14T15:31:31Z","cross_cats_sorted":[],"title_canon_sha256":"b0d75d25cf32a498adb46b18b5b4c6050510265b604e1f24be56a328c2c9de88","abstract_canon_sha256":"1bfe8bbd3a73debcb38d2c1e752b45cad5d44e99250cf9a15442c36a402ec375"},"schema_version":"1.0"},"canonical_sha256":"5b8e6a7417f84ecb0bacbdacc56518d2a4deccea0d5c9e7a1e718519261c9961","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:36:50.655488Z","signature_b64":"M1+LKwqC/4esQq8id3Yx1Rgmor1xqbWErCQtpXNURDgSCukTlJ/vaVgMhj5lSszbZYgivKUYuDQrGtZFmAnNBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5b8e6a7417f84ecb0bacbdacc56518d2a4deccea0d5c9e7a1e718519261c9961","last_reissued_at":"2026-07-05T11:36:50.654901Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:36:50.654901Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2507.10392","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":"2026-07-05T11:36:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DJht/bP+F050gKzr03qOBOU6/3hx+KzWXeGLN9/13WxBRdgSi8sk4XsA6rseNBs5T37fHu7PKMcCiDItecEsDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T04:40:38.473143Z"},"content_sha256":"1ff525bfb937b30982fdfe8734145d10c38b50b71c424591e3f6cf0df82e5dbc","schema_version":"1.0","event_id":"sha256:1ff525bfb937b30982fdfe8734145d10c38b50b71c424591e3f6cf0df82e5dbc"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:LOHGU5AX7BHMWC5MXWWMKZIY2K","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Zorse: Optimizing LLM Training Efficiency on Heterogeneous GPU Clusters","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Khuzaima Daudjee, Rathijit Sen, Runsheng Benson Guo, Utkarsh Anand","submitted_at":"2025-07-14T15:31:31Z","abstract_excerpt":"Large language models (LLMs) require vast amounts of GPU compute to train, but limited availability and high costs of GPUs make homogeneous clusters impractical for many organizations. Instead, assembling heterogeneous clusters by pooling together GPUs of different generations allows them to achieve higher aggregate compute and make use of all available GPUs. However, training on heterogeneous clusters presents several challenges, including load balancing across GPUs, optimizing memory usage to accommodate varying memory capacities, and ensuring communication-efficient training over diverse ne"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.10392","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/2507.10392/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":"2026-07-05T11:36:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"R+FxaJFPkiUQlc67pwgg4S7HBEbLu/R9JPx7BG20VEFLc4M7M2xTitr5aBAXtBfvjKkL53f30KYkonHni+9aCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T04:40:38.473550Z"},"content_sha256":"6b24898b6dd77a87949f6ae6be4055b1ae0543bc36b903569fa3722986e596cc","schema_version":"1.0","event_id":"sha256:6b24898b6dd77a87949f6ae6be4055b1ae0543bc36b903569fa3722986e596cc"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LOHGU5AX7BHMWC5MXWWMKZIY2K/bundle.json","state_url":"https://pith.science/pith/LOHGU5AX7BHMWC5MXWWMKZIY2K/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LOHGU5AX7BHMWC5MXWWMKZIY2K/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":"2026-07-07T04:40:38Z","links":{"resolver":"https://pith.science/pith/LOHGU5AX7BHMWC5MXWWMKZIY2K","bundle":"https://pith.science/pith/LOHGU5AX7BHMWC5MXWWMKZIY2K/bundle.json","state":"https://pith.science/pith/LOHGU5AX7BHMWC5MXWWMKZIY2K/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LOHGU5AX7BHMWC5MXWWMKZIY2K/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:LOHGU5AX7BHMWC5MXWWMKZIY2K","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":"1bfe8bbd3a73debcb38d2c1e752b45cad5d44e99250cf9a15442c36a402ec375","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.DC","submitted_at":"2025-07-14T15:31:31Z","title_canon_sha256":"b0d75d25cf32a498adb46b18b5b4c6050510265b604e1f24be56a328c2c9de88"},"schema_version":"1.0","source":{"id":"2507.10392","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.10392","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"arxiv_version","alias_value":"2507.10392v1","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.10392","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"pith_short_12","alias_value":"LOHGU5AX7BHM","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"pith_short_16","alias_value":"LOHGU5AX7BHMWC5M","created_at":"2026-07-05T11:36:50Z"},{"alias_kind":"pith_short_8","alias_value":"LOHGU5AX","created_at":"2026-07-05T11:36:50Z"}],"graph_snapshots":[{"event_id":"sha256:6b24898b6dd77a87949f6ae6be4055b1ae0543bc36b903569fa3722986e596cc","target":"graph","created_at":"2026-07-05T11:36:50Z","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/2507.10392/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) require vast amounts of GPU compute to train, but limited availability and high costs of GPUs make homogeneous clusters impractical for many organizations. Instead, assembling heterogeneous clusters by pooling together GPUs of different generations allows them to achieve higher aggregate compute and make use of all available GPUs. However, training on heterogeneous clusters presents several challenges, including load balancing across GPUs, optimizing memory usage to accommodate varying memory capacities, and ensuring communication-efficient training over diverse ne","authors_text":"Khuzaima Daudjee, Rathijit Sen, Runsheng Benson Guo, Utkarsh Anand","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.DC","submitted_at":"2025-07-14T15:31:31Z","title":"Zorse: Optimizing LLM Training Efficiency on Heterogeneous GPU Clusters"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.10392","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":"sha256:1ff525bfb937b30982fdfe8734145d10c38b50b71c424591e3f6cf0df82e5dbc","target":"record","created_at":"2026-07-05T11:36:50Z","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":"1bfe8bbd3a73debcb38d2c1e752b45cad5d44e99250cf9a15442c36a402ec375","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.DC","submitted_at":"2025-07-14T15:31:31Z","title_canon_sha256":"b0d75d25cf32a498adb46b18b5b4c6050510265b604e1f24be56a328c2c9de88"},"schema_version":"1.0","source":{"id":"2507.10392","kind":"arxiv","version":1}},"canonical_sha256":"5b8e6a7417f84ecb0bacbdacc56518d2a4deccea0d5c9e7a1e718519261c9961","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5b8e6a7417f84ecb0bacbdacc56518d2a4deccea0d5c9e7a1e718519261c9961","first_computed_at":"2026-07-05T11:36:50.654901Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:36:50.654901Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"M1+LKwqC/4esQq8id3Yx1Rgmor1xqbWErCQtpXNURDgSCukTlJ/vaVgMhj5lSszbZYgivKUYuDQrGtZFmAnNBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:36:50.655488Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.10392","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1ff525bfb937b30982fdfe8734145d10c38b50b71c424591e3f6cf0df82e5dbc","sha256:6b24898b6dd77a87949f6ae6be4055b1ae0543bc36b903569fa3722986e596cc"],"state_sha256":"631ba1046612caaeb6d3800f9743d77410aaaf975de52cea266e5a1817c42b55"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"c4l7KZuCc5+W6fs/pDshQJ2OS1vIkU7lCAi5czaQO1be7MJ6IcdWhZB+fsHIKL6cXua82gCSRcFIVE6gxRxwDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T04:40:38.475786Z","bundle_sha256":"82b253c154019bda8475e7839051e40949428705757aecafe7f5b43dcef50111"}}