{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:LVSFEATXU2QZEJTWUDTAEZNY5D","short_pith_number":"pith:LVSFEATX","canonical_record":{"source":{"id":"2605.12445","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.PF","submitted_at":"2026-05-12T17:39:24Z","cross_cats_sorted":[],"title_canon_sha256":"13fe46a5738069243f192a56cf12f8a2f0cb5ef0791621d6b0f81434930c7572","abstract_canon_sha256":"4712656a39dbfdbb863cc365fd31939af29442d16c40b36e4fd3beed1d1ec4a8"},"schema_version":"1.0"},"canonical_sha256":"5d64520277a6a1922676a0e60265b8e8ee2d9813c6e5781fcd358295093df4cb","source":{"kind":"arxiv","id":"2605.12445","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.12445","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"arxiv_version","alias_value":"2605.12445v2","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12445","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"pith_short_12","alias_value":"LVSFEATXU2QZ","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"pith_short_16","alias_value":"LVSFEATXU2QZEJTW","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"pith_short_8","alias_value":"LVSFEATX","created_at":"2026-05-20T00:05:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:LVSFEATXU2QZEJTWUDTAEZNY5D","target":"record","payload":{"canonical_record":{"source":{"id":"2605.12445","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.PF","submitted_at":"2026-05-12T17:39:24Z","cross_cats_sorted":[],"title_canon_sha256":"13fe46a5738069243f192a56cf12f8a2f0cb5ef0791621d6b0f81434930c7572","abstract_canon_sha256":"4712656a39dbfdbb863cc365fd31939af29442d16c40b36e4fd3beed1d1ec4a8"},"schema_version":"1.0"},"canonical_sha256":"5d64520277a6a1922676a0e60265b8e8ee2d9813c6e5781fcd358295093df4cb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:05:47.344097Z","signature_b64":"O5svM+rtnUuOfzSe7EQxgI26Izk1h/3lMV46ZLYNdgzUb186Zian5GD/ARJ/A4cRLt8nMn3FHfjs/qINKKysBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5d64520277a6a1922676a0e60265b8e8ee2d9813c6e5781fcd358295093df4cb","last_reissued_at":"2026-05-20T00:05:47.343311Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:05:47.343311Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.12445","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":"2026-05-20T00:05:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rbnUhT4Fr1BNfh8cwiNweyKfK2ufD0O2gPe+asrO7f+jWi07wYzzzItjt+LEerZfFJVg9sRRhok2HQTQZxhZDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-01T05:57:02.105674Z"},"content_sha256":"6f95deeca9ccee0022b7eec451da05e2e3a955017e9339ea763d722c151a21d4","schema_version":"1.0","event_id":"sha256:6f95deeca9ccee0022b7eec451da05e2e3a955017e9339ea763d722c151a21d4"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:LVSFEATXU2QZEJTWUDTAEZNY5D","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Scalable Packed Layouts for Vector-Length-Agnostic ML Code Generation","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"Vector-length-aware packed data layouts enable ML compilers to generate efficient vector-length-agnostic code for scalable vector hardware.","cross_cats":[],"primary_cat":"cs.PF","authors_text":"Ege Beysel, Jan Moritz Joseph, Maximilian Bartel","submitted_at":"2026-05-12T17:39:24Z","abstract_excerpt":"Scalable vector instruction sets such as Arm SVE enable vector-length-agnostic (VLA) execution, allowing a single implementation to adapt across hardware with different vector lengths. However, they complicate compiler code generation, as tiling and data layout decisions can no longer be fixed at compile time.\n  We present an approach for enabling VLA code generation in an end-to-end ML compilation pipeline through vector-length-aware packed data layouts and corresponding compiler extensions. We integrate these mechanisms into MLIR/IREE and extend tiling, fusion, and vectorization to operate w"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We present an approach for enabling VLA code generation in an end-to-end ML compilation pipeline through vector-length-aware packed data layouts and corresponding compiler extensions. ... Evaluated on real-world ML workloads on Arm CPUs, our approach generates SVE code that is competitive with, and often outperforms, existing NEON-based code generation within IREE, achieving up to 1.45× speedup.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the vector-length-aware packed data layouts and extensions to tiling, fusion, and vectorization can be integrated into MLIR/IREE without introducing correctness issues, significant runtime overhead, or suboptimal performance for untested workloads, and that the evaluated ML tasks represent general cases.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Packed layouts and extensions to tiling/fusion/vectorization in MLIR/IREE enable VLA ML code generation for SVE, achieving up to 1.45x speedup over NEON and outperforming PyTorch frameworks while scaling with vector length.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Vector-length-aware packed data layouts enable ML compilers to generate efficient vector-length-agnostic code for scalable vector hardware.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9e64935e87ef59bd2e80c7e08daa762761deb8aeb919363a34ae919df60c9f24"},"source":{"id":"2605.12445","kind":"arxiv","version":2},"verdict":{"id":"b7e6e3d0-9e7a-46a0-93e2-85d18d915d5d","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-13T02:16:38.666513Z","strongest_claim":"We present an approach for enabling VLA code generation in an end-to-end ML compilation pipeline through vector-length-aware packed data layouts and corresponding compiler extensions. ... Evaluated on real-world ML workloads on Arm CPUs, our approach generates SVE code that is competitive with, and often outperforms, existing NEON-based code generation within IREE, achieving up to 1.45× speedup.","one_line_summary":"Packed layouts and extensions to tiling/fusion/vectorization in MLIR/IREE enable VLA ML code generation for SVE, achieving up to 1.45x speedup over NEON and outperforming PyTorch frameworks while scaling with vector length.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the vector-length-aware packed data layouts and extensions to tiling, fusion, and vectorization can be integrated into MLIR/IREE without introducing correctness issues, significant runtime overhead, or suboptimal performance for untested workloads, and that the evaluated ML tasks represent general cases.","pith_extraction_headline":"Vector-length-aware packed data layouts enable ML compilers to generate efficient vector-length-agnostic code for scalable vector hardware."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.12445/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-19T22:21:57.834723Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T10:35:08.813860Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T08:01:18.277438Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T07:27:19.644305Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"a597cb7ed85303185c56a7c28c7f45dcc6290942051c2ea0a91cad00346690f3"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"216e7112f674b24d0ed3912194d547ae17ee9b6ff33410700b0b20261a686a62"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"b7e6e3d0-9e7a-46a0-93e2-85d18d915d5d"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:05:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ugv4EOEz+4tZclYzOdvZKeMuxLvj2aXzPvwONkOBaUdU3/fn9GtEMxQdI2s/8bjBCbaRFr/HE586es8EfgDFBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-01T05:57:02.106238Z"},"content_sha256":"1c13dab0265c816b5f97bb94ea5e237e0a4e7711ae758980204fde39458fb3b3","schema_version":"1.0","event_id":"sha256:1c13dab0265c816b5f97bb94ea5e237e0a4e7711ae758980204fde39458fb3b3"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LVSFEATXU2QZEJTWUDTAEZNY5D/bundle.json","state_url":"https://pith.science/pith/LVSFEATXU2QZEJTWUDTAEZNY5D/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LVSFEATXU2QZEJTWUDTAEZNY5D/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":"2026-06-01T05:57:02Z","links":{"resolver":"https://pith.science/pith/LVSFEATXU2QZEJTWUDTAEZNY5D","bundle":"https://pith.science/pith/LVSFEATXU2QZEJTWUDTAEZNY5D/bundle.json","state":"https://pith.science/pith/LVSFEATXU2QZEJTWUDTAEZNY5D/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LVSFEATXU2QZEJTWUDTAEZNY5D/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:LVSFEATXU2QZEJTWUDTAEZNY5D","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":"4712656a39dbfdbb863cc365fd31939af29442d16c40b36e4fd3beed1d1ec4a8","cross_cats_sorted":[],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.PF","submitted_at":"2026-05-12T17:39:24Z","title_canon_sha256":"13fe46a5738069243f192a56cf12f8a2f0cb5ef0791621d6b0f81434930c7572"},"schema_version":"1.0","source":{"id":"2605.12445","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.12445","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"arxiv_version","alias_value":"2605.12445v2","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12445","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"pith_short_12","alias_value":"LVSFEATXU2QZ","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"pith_short_16","alias_value":"LVSFEATXU2QZEJTW","created_at":"2026-05-20T00:05:47Z"},{"alias_kind":"pith_short_8","alias_value":"LVSFEATX","created_at":"2026-05-20T00:05:47Z"}],"graph_snapshots":[{"event_id":"sha256:1c13dab0265c816b5f97bb94ea5e237e0a4e7711ae758980204fde39458fb3b3","target":"graph","created_at":"2026-05-20T00:05:47Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"We present an approach for enabling VLA code generation in an end-to-end ML compilation pipeline through vector-length-aware packed data layouts and corresponding compiler extensions. ... Evaluated on real-world ML workloads on Arm CPUs, our approach generates SVE code that is competitive with, and often outperforms, existing NEON-based code generation within IREE, achieving up to 1.45× speedup."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the vector-length-aware packed data layouts and extensions to tiling, fusion, and vectorization can be integrated into MLIR/IREE without introducing correctness issues, significant runtime overhead, or suboptimal performance for untested workloads, and that the evaluated ML tasks represent general cases."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Packed layouts and extensions to tiling/fusion/vectorization in MLIR/IREE enable VLA ML code generation for SVE, achieving up to 1.45x speedup over NEON and outperforming PyTorch frameworks while scaling with vector length."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Vector-length-aware packed data layouts enable ML compilers to generate efficient vector-length-agnostic code for scalable vector hardware."}],"snapshot_sha256":"9e64935e87ef59bd2e80c7e08daa762761deb8aeb919363a34ae919df60c9f24"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"216e7112f674b24d0ed3912194d547ae17ee9b6ff33410700b0b20261a686a62"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T22:21:57.834723Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T10:35:08.813860Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T08:01:18.277438Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T07:27:19.644305Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.12445/integrity.json","findings":[],"snapshot_sha256":"a597cb7ed85303185c56a7c28c7f45dcc6290942051c2ea0a91cad00346690f3","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Scalable vector instruction sets such as Arm SVE enable vector-length-agnostic (VLA) execution, allowing a single implementation to adapt across hardware with different vector lengths. However, they complicate compiler code generation, as tiling and data layout decisions can no longer be fixed at compile time.\n  We present an approach for enabling VLA code generation in an end-to-end ML compilation pipeline through vector-length-aware packed data layouts and corresponding compiler extensions. We integrate these mechanisms into MLIR/IREE and extend tiling, fusion, and vectorization to operate w","authors_text":"Ege Beysel, Jan Moritz Joseph, Maximilian Bartel","cross_cats":[],"headline":"Vector-length-aware packed data layouts enable ML compilers to generate efficient vector-length-agnostic code for scalable vector hardware.","license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.PF","submitted_at":"2026-05-12T17:39:24Z","title":"Scalable Packed Layouts for Vector-Length-Agnostic ML Code Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.12445","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-13T02:16:38.666513Z","id":"b7e6e3d0-9e7a-46a0-93e2-85d18d915d5d","model_set":{"reader":"grok-4.3"},"one_line_summary":"Packed layouts and extensions to tiling/fusion/vectorization in MLIR/IREE enable VLA ML code generation for SVE, achieving up to 1.45x speedup over NEON and outperforming PyTorch frameworks while scaling with vector length.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Vector-length-aware packed data layouts enable ML compilers to generate efficient vector-length-agnostic code for scalable vector hardware.","strongest_claim":"We present an approach for enabling VLA code generation in an end-to-end ML compilation pipeline through vector-length-aware packed data layouts and corresponding compiler extensions. ... Evaluated on real-world ML workloads on Arm CPUs, our approach generates SVE code that is competitive with, and often outperforms, existing NEON-based code generation within IREE, achieving up to 1.45× speedup.","weakest_assumption":"That the vector-length-aware packed data layouts and extensions to tiling, fusion, and vectorization can be integrated into MLIR/IREE without introducing correctness issues, significant runtime overhead, or suboptimal performance for untested workloads, and that the evaluated ML tasks represent general cases."}},"verdict_id":"b7e6e3d0-9e7a-46a0-93e2-85d18d915d5d"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6f95deeca9ccee0022b7eec451da05e2e3a955017e9339ea763d722c151a21d4","target":"record","created_at":"2026-05-20T00:05:47Z","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":"4712656a39dbfdbb863cc365fd31939af29442d16c40b36e4fd3beed1d1ec4a8","cross_cats_sorted":[],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.PF","submitted_at":"2026-05-12T17:39:24Z","title_canon_sha256":"13fe46a5738069243f192a56cf12f8a2f0cb5ef0791621d6b0f81434930c7572"},"schema_version":"1.0","source":{"id":"2605.12445","kind":"arxiv","version":2}},"canonical_sha256":"5d64520277a6a1922676a0e60265b8e8ee2d9813c6e5781fcd358295093df4cb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5d64520277a6a1922676a0e60265b8e8ee2d9813c6e5781fcd358295093df4cb","first_computed_at":"2026-05-20T00:05:47.343311Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:05:47.343311Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"O5svM+rtnUuOfzSe7EQxgI26Izk1h/3lMV46ZLYNdgzUb186Zian5GD/ARJ/A4cRLt8nMn3FHfjs/qINKKysBQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:05:47.344097Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.12445","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6f95deeca9ccee0022b7eec451da05e2e3a955017e9339ea763d722c151a21d4","sha256:1c13dab0265c816b5f97bb94ea5e237e0a4e7711ae758980204fde39458fb3b3"],"state_sha256":"40be36cca58bdf165bb447186d031581b8c65feb2e11e7a836866ad5ea14c811"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FFMisrlqYBmtrHevFYj1XzzWNI5o7Nqdd63zevJa9//1fmJaD6A6yPzjbgwJieCqOvzSLHqk0Bl8GmcQ0f+XAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-01T05:57:02.108951Z","bundle_sha256":"be1a3cc85c174525cbb2107a96a887a0e3c13d591fec308b96ad0255387618fc"}}