{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:O2Q6VE6IPXBFTEJ2KWZR664MM3","short_pith_number":"pith:O2Q6VE6I","schema_version":"1.0","canonical_sha256":"76a1ea93c87dc259913a55b31f7b8c66c8db14a24243c3af654f71309de5fba4","source":{"kind":"arxiv","id":"2311.09431","version":1},"attestation_state":"computed","paper":{"title":"Striped Attention: Faster Ring Attention for Causal Transformers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Aniruddha Nrusimha, Jonathan Ragan-Kelley, Kevin Qian, Tian Jin, William Brandon, Zachary Ankner, Zhiye Song","submitted_at":"2023-11-15T23:01:02Z","abstract_excerpt":"To help address the growing demand for ever-longer sequence lengths in transformer models, Liu et al. recently proposed Ring Attention, an exact attention algorithm capable of overcoming per-device memory bottle- necks by distributing self-attention across multiple devices. In this paper, we study the performance characteristics of Ring Attention in the important special case of causal transformer models, and identify a key workload imbal- ance due to triangular structure of causal attention computations. We propose a simple extension to Ring Attention, which we call Striped Attention to fix t"},"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":"2311.09431","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-11-15T23:01:02Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"97f51caa7359f692ee229bb3f1704255052b13215bccb09d172eac0b47b4d341","abstract_canon_sha256":"434cf4dcd16600e574c4876310f49613dc4ec2d9b0014666ac3e08199ece6058"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:13:27.911669Z","signature_b64":"QeTeNfTBmhzIepa/w+OaYnXGnZc2aEqBlBEAK1lbnEJKKbfp5giIuULB6F6dNUxP7m+1K2od4di5jgH0xZa7BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"76a1ea93c87dc259913a55b31f7b8c66c8db14a24243c3af654f71309de5fba4","last_reissued_at":"2026-07-05T07:13:27.911045Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:13:27.911045Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Striped Attention: Faster Ring Attention for Causal Transformers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Aniruddha Nrusimha, Jonathan Ragan-Kelley, Kevin Qian, Tian Jin, William Brandon, Zachary Ankner, Zhiye Song","submitted_at":"2023-11-15T23:01:02Z","abstract_excerpt":"To help address the growing demand for ever-longer sequence lengths in transformer models, Liu et al. recently proposed Ring Attention, an exact attention algorithm capable of overcoming per-device memory bottle- necks by distributing self-attention across multiple devices. In this paper, we study the performance characteristics of Ring Attention in the important special case of causal transformer models, and identify a key workload imbal- ance due to triangular structure of causal attention computations. We propose a simple extension to Ring Attention, which we call Striped Attention to fix t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.09431","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/2311.09431/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":"2311.09431","created_at":"2026-07-05T07:13:27.911102+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.09431v1","created_at":"2026-07-05T07:13:27.911102+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.09431","created_at":"2026-07-05T07:13:27.911102+00:00"},{"alias_kind":"pith_short_12","alias_value":"O2Q6VE6IPXBF","created_at":"2026-07-05T07:13:27.911102+00:00"},{"alias_kind":"pith_short_16","alias_value":"O2Q6VE6IPXBFTEJ2","created_at":"2026-07-05T07:13:27.911102+00:00"},{"alias_kind":"pith_short_8","alias_value":"O2Q6VE6I","created_at":"2026-07-05T07:13:27.911102+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":15,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22541","citing_title":"ASAP: A Disaggregated and Asynchronous Inference System for MoE Prefill","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10493","citing_title":"Achieving Cloud-Grade SLOs for Local Mixture-of-Experts Inference through CPU-GPU Hybrid Design","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01701","citing_title":"Arachne: Orchestrating Cascades for Efficient Text-to-Video Model Training","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01817","citing_title":"HCMS: Head-Chunked Multi-Stream Pipeline for Communication-Computation Overlap in Long-Sequence Parallel Attention","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2606.17059","citing_title":"Towards Distributed Inference of LLMs on a P2P Network","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2510.18830","citing_title":"MTraining: Distributed Dynamic Sparse Attention for Efficient Ultra-Long Context Training","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2407.08608","citing_title":"FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precision","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2509.21275","citing_title":"InfiniPipe: Elastic Pipeline Parallelism for Efficient Variable-Length Long-Context LLM Training","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2402.08268","citing_title":"World Model on Million-Length Video And Language With Blockwise RingAttention","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2312.06635","citing_title":"Gated Linear Attention Transformers with Hardware-Efficient Training","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2603.27960","citing_title":"Towards Efficient Large Vision-Language Models: A Comprehensive Survey on Inference Strategies","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08587","citing_title":"Kaczmarz Linear Attention","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2405.21060","citing_title":"Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07569","citing_title":"HexiSeq: Accommodating Long Context Training of LLMs over Heterogeneous Hardware","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14561","citing_title":"CoCoDiff: Optimizing Collective Communications for Distributed Diffusion Transformer Inference Under Ulysses Sequence Parallelism","ref_index":55,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/O2Q6VE6IPXBFTEJ2KWZR664MM3","json":"https://pith.science/pith/O2Q6VE6IPXBFTEJ2KWZR664MM3.json","graph_json":"https://pith.science/api/pith-number/O2Q6VE6IPXBFTEJ2KWZR664MM3/graph.json","events_json":"https://pith.science/api/pith-number/O2Q6VE6IPXBFTEJ2KWZR664MM3/events.json","paper":"https://pith.science/paper/O2Q6VE6I"},"agent_actions":{"view_html":"https://pith.science/pith/O2Q6VE6IPXBFTEJ2KWZR664MM3","download_json":"https://pith.science/pith/O2Q6VE6IPXBFTEJ2KWZR664MM3.json","view_paper":"https://pith.science/paper/O2Q6VE6I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.09431&json=true","fetch_graph":"https://pith.science/api/pith-number/O2Q6VE6IPXBFTEJ2KWZR664MM3/graph.json","fetch_events":"https://pith.science/api/pith-number/O2Q6VE6IPXBFTEJ2KWZR664MM3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O2Q6VE6IPXBFTEJ2KWZR664MM3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O2Q6VE6IPXBFTEJ2KWZR664MM3/action/storage_attestation","attest_author":"https://pith.science/pith/O2Q6VE6IPXBFTEJ2KWZR664MM3/action/author_attestation","sign_citation":"https://pith.science/pith/O2Q6VE6IPXBFTEJ2KWZR664MM3/action/citation_signature","submit_replication":"https://pith.science/pith/O2Q6VE6IPXBFTEJ2KWZR664MM3/action/replication_record"}},"created_at":"2026-07-05T07:13:27.911102+00:00","updated_at":"2026-07-05T07:13:27.911102+00:00"}