{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:X57PITENTCPLE4YKNEUENFRBIZ","short_pith_number":"pith:X57PITEN","schema_version":"1.0","canonical_sha256":"bf7ef44c8d989eb2730a69284696214673d846d206dcd5bf9293d4e8ee32cf98","source":{"kind":"arxiv","id":"2006.10901","version":2},"attestation_state":"computed","paper":{"title":"Sparse GPU Kernels for Deep Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Cliff Young, Erich Elsen, Matei Zaharia, Trevor Gale","submitted_at":"2020-06-18T23:59:11Z","abstract_excerpt":"Scientific workloads have traditionally exploited high levels of sparsity to accelerate computation and reduce memory requirements. While deep neural networks can be made sparse, achieving practical speedups on GPUs is difficult because these applications have relatively moderate levels of sparsity that are not sufficient for existing sparse kernels to outperform their dense counterparts. In this work, we study sparse matrices from deep learning applications and identify favorable properties that can be exploited to accelerate computation. Based on these insights, we develop high-performance G"},"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":"2006.10901","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-06-18T23:59:11Z","cross_cats_sorted":["cs.DC","stat.ML"],"title_canon_sha256":"da321d09e971adcfddcdba1f44660388f5e8838fc3edd4388e8a9a70352eb0d7","abstract_canon_sha256":"fc6dec74f25986505c0401330d1e3b87ccbf31cfafc04184d7f08f29e241018f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:32:08.385417Z","signature_b64":"0n9/+nf3QDDZtxw4n/eVmXlqHW4zszXyTmloQug1wiCUW/DVPvCx98EoJVGxmzVcZI2SqXWiuJLzp/ySbnBAAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bf7ef44c8d989eb2730a69284696214673d846d206dcd5bf9293d4e8ee32cf98","last_reissued_at":"2026-07-05T01:32:08.384916Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:32:08.384916Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sparse GPU Kernels for Deep Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Cliff Young, Erich Elsen, Matei Zaharia, Trevor Gale","submitted_at":"2020-06-18T23:59:11Z","abstract_excerpt":"Scientific workloads have traditionally exploited high levels of sparsity to accelerate computation and reduce memory requirements. While deep neural networks can be made sparse, achieving practical speedups on GPUs is difficult because these applications have relatively moderate levels of sparsity that are not sufficient for existing sparse kernels to outperform their dense counterparts. In this work, we study sparse matrices from deep learning applications and identify favorable properties that can be exploited to accelerate computation. Based on these insights, we develop high-performance G"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.10901","kind":"arxiv","version":2},"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/2006.10901/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":"2006.10901","created_at":"2026-07-05T01:32:08.384975+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.10901v2","created_at":"2026-07-05T01:32:08.384975+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.10901","created_at":"2026-07-05T01:32:08.384975+00:00"},{"alias_kind":"pith_short_12","alias_value":"X57PITENTCPL","created_at":"2026-07-05T01:32:08.384975+00:00"},{"alias_kind":"pith_short_16","alias_value":"X57PITENTCPLE4YK","created_at":"2026-07-05T01:32:08.384975+00:00"},{"alias_kind":"pith_short_8","alias_value":"X57PITEN","created_at":"2026-07-05T01:32:08.384975+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22482","citing_title":"NeutronSparse: Coordinating Heterogeneous Engines for Sparse Matrix Multiplication on NPUs","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2101.03961","citing_title":"Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2202.08906","citing_title":"ST-MoE: Designing Stable and Transferable Sparse Expert Models","ref_index":69,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/X57PITENTCPLE4YKNEUENFRBIZ","json":"https://pith.science/pith/X57PITENTCPLE4YKNEUENFRBIZ.json","graph_json":"https://pith.science/api/pith-number/X57PITENTCPLE4YKNEUENFRBIZ/graph.json","events_json":"https://pith.science/api/pith-number/X57PITENTCPLE4YKNEUENFRBIZ/events.json","paper":"https://pith.science/paper/X57PITEN"},"agent_actions":{"view_html":"https://pith.science/pith/X57PITENTCPLE4YKNEUENFRBIZ","download_json":"https://pith.science/pith/X57PITENTCPLE4YKNEUENFRBIZ.json","view_paper":"https://pith.science/paper/X57PITEN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.10901&json=true","fetch_graph":"https://pith.science/api/pith-number/X57PITENTCPLE4YKNEUENFRBIZ/graph.json","fetch_events":"https://pith.science/api/pith-number/X57PITENTCPLE4YKNEUENFRBIZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X57PITENTCPLE4YKNEUENFRBIZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X57PITENTCPLE4YKNEUENFRBIZ/action/storage_attestation","attest_author":"https://pith.science/pith/X57PITENTCPLE4YKNEUENFRBIZ/action/author_attestation","sign_citation":"https://pith.science/pith/X57PITENTCPLE4YKNEUENFRBIZ/action/citation_signature","submit_replication":"https://pith.science/pith/X57PITENTCPLE4YKNEUENFRBIZ/action/replication_record"}},"created_at":"2026-07-05T01:32:08.384975+00:00","updated_at":"2026-07-05T01:32:08.384975+00:00"}