{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:5YLGAVGW73KT7L3U2XU5CCYRDP","short_pith_number":"pith:5YLGAVGW","schema_version":"1.0","canonical_sha256":"ee166054d6fed53faf74d5e9d10b111bc1d9aae2df0fa76e1528839d66af54db","source":{"kind":"arxiv","id":"2101.05615","version":1},"attestation_state":"computed","paper":{"title":"FBGEMM: Enabling High-Performance Low-Precision Deep Learning Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.PF"],"primary_cat":"cs.LG","authors_text":"Daya Khudia, Haixin Liu, Jianyu Huang, Jongsoo Park, Mikhail Smelyanskiy, Protonu Basu, Summer Deng","submitted_at":"2021-01-13T00:34:04Z","abstract_excerpt":"Deep learning models typically use single-precision (FP32) floating point data types for representing activations and weights, but a slew of recent research work has shown that computations with reduced-precision data types (FP16, 16-bit integers, 8-bit integers or even 4- or 2-bit integers) are enough to achieve same accuracy as FP32 and are much more efficient. Therefore, we designed fbgemm, a high-performance kernel library, from ground up to perform high-performance quantized inference on current generation CPUs. fbgemm achieves efficiency by fusing common quantization operations with a hi"},"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":"2101.05615","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-01-13T00:34:04Z","cross_cats_sorted":["cs.PF"],"title_canon_sha256":"de471a0711565a4e5edbe71331fc1bc17fb2b78b432ab4cd9e4622500c0f3512","abstract_canon_sha256":"2d22d613195503f17aa74909a9c2ddb67ed79f06a442dd4aed4809fe45db6430"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:06:57.528091Z","signature_b64":"HIr2f7t4iNfaatFgr4mbbfaHd5TIZMLjtub5IJf5oCIRq1qL0A9TCig/BHVAWiZVhl5NCHt5xr1ULIECWfRbDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ee166054d6fed53faf74d5e9d10b111bc1d9aae2df0fa76e1528839d66af54db","last_reissued_at":"2026-07-05T02:06:57.527639Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:06:57.527639Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FBGEMM: Enabling High-Performance Low-Precision Deep Learning Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.PF"],"primary_cat":"cs.LG","authors_text":"Daya Khudia, Haixin Liu, Jianyu Huang, Jongsoo Park, Mikhail Smelyanskiy, Protonu Basu, Summer Deng","submitted_at":"2021-01-13T00:34:04Z","abstract_excerpt":"Deep learning models typically use single-precision (FP32) floating point data types for representing activations and weights, but a slew of recent research work has shown that computations with reduced-precision data types (FP16, 16-bit integers, 8-bit integers or even 4- or 2-bit integers) are enough to achieve same accuracy as FP32 and are much more efficient. Therefore, we designed fbgemm, a high-performance kernel library, from ground up to perform high-performance quantized inference on current generation CPUs. fbgemm achieves efficiency by fusing common quantization operations with a hi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2101.05615","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/2101.05615/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":"2101.05615","created_at":"2026-07-05T02:06:57.527710+00:00"},{"alias_kind":"arxiv_version","alias_value":"2101.05615v1","created_at":"2026-07-05T02:06:57.527710+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2101.05615","created_at":"2026-07-05T02:06:57.527710+00:00"},{"alias_kind":"pith_short_12","alias_value":"5YLGAVGW73KT","created_at":"2026-07-05T02:06:57.527710+00:00"},{"alias_kind":"pith_short_16","alias_value":"5YLGAVGW73KT7L3U","created_at":"2026-07-05T02:06:57.527710+00:00"},{"alias_kind":"pith_short_8","alias_value":"5YLGAVGW","created_at":"2026-07-05T02:06:57.527710+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":11,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25562","citing_title":"Energy-Efficient CNN Acceleration with MSDF Digit-Serial Arithmetic on FPGA","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21101","citing_title":"DPIFrame: A Dual-Level Parallelism Acceleration Framework for CTR Model Inference","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2405.13170","citing_title":"FEATHER: A Reconfigurable Accelerator with Data Reordering Support for Low-Cost On-Chip Dataflow Switching","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10886","citing_title":"LoKA: Low-precision Kernel Applications for Recommendation Models At Scale","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2402.17152","citing_title":"Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations","ref_index":124,"is_internal_anchor":false},{"citing_arxiv_id":"2208.07339","citing_title":"LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scale","ref_index":142,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10886","citing_title":"LoKA: Low-precision Kernel Applications for Recommendation Models At Scale","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25799","citing_title":"At the Edge of the Heart: ULP FPGA-Based CNN for On-Device Cardiac Feature Extraction in Smart Health Sensors for Astronauts","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07719","citing_title":"An Efficient Hybrid Sparse Attention with CPU-GPU Parallelism for Long-Context Inference","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17476","citing_title":"Privatar: Scalable Privacy-preserving Multi-user VR via Secure Offloading","ref_index":160,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04450","citing_title":"One Pool, Two Caches: Adaptive HBM Partitioning for Accelerating Generative Recommender Serving","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5YLGAVGW73KT7L3U2XU5CCYRDP","json":"https://pith.science/pith/5YLGAVGW73KT7L3U2XU5CCYRDP.json","graph_json":"https://pith.science/api/pith-number/5YLGAVGW73KT7L3U2XU5CCYRDP/graph.json","events_json":"https://pith.science/api/pith-number/5YLGAVGW73KT7L3U2XU5CCYRDP/events.json","paper":"https://pith.science/paper/5YLGAVGW"},"agent_actions":{"view_html":"https://pith.science/pith/5YLGAVGW73KT7L3U2XU5CCYRDP","download_json":"https://pith.science/pith/5YLGAVGW73KT7L3U2XU5CCYRDP.json","view_paper":"https://pith.science/paper/5YLGAVGW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2101.05615&json=true","fetch_graph":"https://pith.science/api/pith-number/5YLGAVGW73KT7L3U2XU5CCYRDP/graph.json","fetch_events":"https://pith.science/api/pith-number/5YLGAVGW73KT7L3U2XU5CCYRDP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5YLGAVGW73KT7L3U2XU5CCYRDP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5YLGAVGW73KT7L3U2XU5CCYRDP/action/storage_attestation","attest_author":"https://pith.science/pith/5YLGAVGW73KT7L3U2XU5CCYRDP/action/author_attestation","sign_citation":"https://pith.science/pith/5YLGAVGW73KT7L3U2XU5CCYRDP/action/citation_signature","submit_replication":"https://pith.science/pith/5YLGAVGW73KT7L3U2XU5CCYRDP/action/replication_record"}},"created_at":"2026-07-05T02:06:57.527710+00:00","updated_at":"2026-07-05T02:06:57.527710+00:00"}