{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:LH46NL4RHDJA3SOIXY4DRFBKKT","short_pith_number":"pith:LH46NL4R","schema_version":"1.0","canonical_sha256":"59f9e6af9138d20dc9c8be3838942a54eb7bed2192882348f7071eba02cbdd4c","source":{"kind":"arxiv","id":"2309.07062","version":1},"attestation_state":"computed","paper":{"title":"Large Language Models for Compiler Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.PL","authors_text":"Baptiste Roziere, Chris Cummins, Dejan Grubisic, Fabian Gloeckle, Gabriel Synnaeve, Hugh Leather, Jonas Gehring, Kim Hazelwood, Mostafa Elhoushi, Volker Seeker, Youwei Liang","submitted_at":"2023-09-11T22:11:46Z","abstract_excerpt":"We explore the novel application of Large Language Models to code optimization. We present a 7B-parameter transformer model trained from scratch to optimize LLVM assembly for code size. The model takes as input unoptimized assembly and outputs a list of compiler options to best optimize the program. Crucially, during training, we ask the model to predict the instruction counts before and after optimization, and the optimized code itself. These auxiliary learning tasks significantly improve the optimization performance of the model and improve the model's depth of understanding.\n  We evaluate o"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2309.07062","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.PL","submitted_at":"2023-09-11T22:11:46Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"1ba19613a9f96821379ec444f22f346f0aa61e67c33da58cb9349c95ea601e80","abstract_canon_sha256":"e1605a6f6de987ca2c27f1b7d1b4c1d28bf22bde942e88c1cc537862068f7023"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:50:25.769865Z","signature_b64":"0X5cUCW39Icws8Wt+LEoQvZMoEF99XdzjoHXHeQPggJd4Es5jYPnw4xBn2tNmtmauRCytKc8Q9PO4CJzZe9TBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"59f9e6af9138d20dc9c8be3838942a54eb7bed2192882348f7071eba02cbdd4c","last_reissued_at":"2026-07-05T06:50:25.769372Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:50:25.769372Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Large Language Models for Compiler Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.PL","authors_text":"Baptiste Roziere, Chris Cummins, Dejan Grubisic, Fabian Gloeckle, Gabriel Synnaeve, Hugh Leather, Jonas Gehring, Kim Hazelwood, Mostafa Elhoushi, Volker Seeker, Youwei Liang","submitted_at":"2023-09-11T22:11:46Z","abstract_excerpt":"We explore the novel application of Large Language Models to code optimization. We present a 7B-parameter transformer model trained from scratch to optimize LLVM assembly for code size. The model takes as input unoptimized assembly and outputs a list of compiler options to best optimize the program. Crucially, during training, we ask the model to predict the instruction counts before and after optimization, and the optimized code itself. These auxiliary learning tasks significantly improve the optimization performance of the model and improve the model's depth of understanding.\n  We evaluate o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.07062","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/2309.07062/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2309.07062","created_at":"2026-07-05T06:50:25.769430+00:00"},{"alias_kind":"arxiv_version","alias_value":"2309.07062v1","created_at":"2026-07-05T06:50:25.769430+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.07062","created_at":"2026-07-05T06:50:25.769430+00:00"},{"alias_kind":"pith_short_12","alias_value":"LH46NL4RHDJA","created_at":"2026-07-05T06:50:25.769430+00:00"},{"alias_kind":"pith_short_16","alias_value":"LH46NL4RHDJA3SOI","created_at":"2026-07-05T06:50:25.769430+00:00"},{"alias_kind":"pith_short_8","alias_value":"LH46NL4R","created_at":"2026-07-05T06:50:25.769430+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23512","citing_title":"Source-Free Detection and Impact Analysis of Compiler Optimization Problems in Mobile Applications","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07665","citing_title":"AgentCompile: An LLM-Guided Compiler for Direct CUDA Inference","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05570","citing_title":"TensorBench: Benchmarking Coding Agents on a Compiler-Based Tensor Framework","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31767","citing_title":"JETO-Bench: A Reproducible Benchmark for Execution Time Improvement Patches in Java","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25954","citing_title":"Step-TP: A Grounded, Step-Level Dataset with Chain-of-Thought Reasoning for LLM-Guided Tensor Program Optimization","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22356","citing_title":"Modeling Pathology-Like Behavioral Patterns in Language Models Through Behavioral Fine-Tuning","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03144","citing_title":"InCoder-32B-Thinking: Industrial Code World Model for Thinking","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08247","citing_title":"LLM Translation of Compiler Intermediate Representation","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LH46NL4RHDJA3SOIXY4DRFBKKT","json":"https://pith.science/pith/LH46NL4RHDJA3SOIXY4DRFBKKT.json","graph_json":"https://pith.science/api/pith-number/LH46NL4RHDJA3SOIXY4DRFBKKT/graph.json","events_json":"https://pith.science/api/pith-number/LH46NL4RHDJA3SOIXY4DRFBKKT/events.json","paper":"https://pith.science/paper/LH46NL4R"},"agent_actions":{"view_html":"https://pith.science/pith/LH46NL4RHDJA3SOIXY4DRFBKKT","download_json":"https://pith.science/pith/LH46NL4RHDJA3SOIXY4DRFBKKT.json","view_paper":"https://pith.science/paper/LH46NL4R","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2309.07062&json=true","fetch_graph":"https://pith.science/api/pith-number/LH46NL4RHDJA3SOIXY4DRFBKKT/graph.json","fetch_events":"https://pith.science/api/pith-number/LH46NL4RHDJA3SOIXY4DRFBKKT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LH46NL4RHDJA3SOIXY4DRFBKKT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LH46NL4RHDJA3SOIXY4DRFBKKT/action/storage_attestation","attest_author":"https://pith.science/pith/LH46NL4RHDJA3SOIXY4DRFBKKT/action/author_attestation","sign_citation":"https://pith.science/pith/LH46NL4RHDJA3SOIXY4DRFBKKT/action/citation_signature","submit_replication":"https://pith.science/pith/LH46NL4RHDJA3SOIXY4DRFBKKT/action/replication_record"}},"created_at":"2026-07-05T06:50:25.769430+00:00","updated_at":"2026-07-05T06:50:25.769430+00:00"}