{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,11]],"date-time":"2025-12-11T07:41:03Z","timestamp":1765438863346,"version":"3.37.3"},"reference-count":58,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"6","license":[{"start":{"date-parts":[[2023,6,1]],"date-time":"2023-06-01T00:00:00Z","timestamp":1685577600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2023,6,1]],"date-time":"2023-06-01T00:00:00Z","timestamp":1685577600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2023,6,1]],"date-time":"2023-06-01T00:00:00Z","timestamp":1685577600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001381","name":"National Research Foundation Singapore","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001381","id-type":"DOI","asserted-by":"publisher"}]},{"name":"AI Singapore Programme","award":["AISG2-TC-2021-002"],"award-info":[{"award-number":["AISG2-TC-2021-002"]}]},{"name":"Ministry of Education AcRF Tier 2","award":["MOE-000242-00","MOE-000242-01"],"award-info":[{"award-number":["MOE-000242-00","MOE-000242-01"]}]},{"name":"Centre for Trusted Internet and Community"},{"DOI":"10.13039\/501100001352","name":"National University of Singapore","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001352","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Parallel Distrib. Syst."],"published-print":{"date-parts":[[2023,6]]},"DOI":"10.1109\/tpds.2023.3269530","type":"journal-article","created":{"date-parts":[[2023,4,24]],"date-time":"2023-04-24T18:38:54Z","timestamp":1682361534000},"page":"1982-1995","source":"Crossref","is-referenced-by-count":7,"title":["NIOT: A Novel Inference Optimization of Transformers on Modern CPUs"],"prefix":"10.1109","volume":"34","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-7992-426X","authenticated-orcid":false,"given":"Zining","family":"Zhang","sequence":"first","affiliation":[{"name":"School of Computing, National University of Singapore, Singapore"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5798-2282","authenticated-orcid":false,"given":"Yao","family":"Chen","sequence":"additional","affiliation":[{"name":"School of Computing, National University of Singapore, Singapore"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8618-4581","authenticated-orcid":false,"given":"Bingsheng","family":"He","sequence":"additional","affiliation":[{"name":"School of Computing, National University of Singapore, Singapore"}]},{"given":"Zhenjie","family":"Zhang","sequence":"additional","affiliation":[{"name":"Neuron Mobility Pte. Ltd., Singapore"}]}],"member":"263","reference":[{"article-title":"Efficient 8-bit quantization of transformer neural machine language translation model","year":"2019","author":"bhandare","key":"ref13"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA53966.2022.00064"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.emnlp-main.211"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP39728.2021.9413670"},{"article-title":"Generating long sequences with sparse transformers","year":"2019","author":"child","key":"ref15"},{"key":"ref14","first-page":"3393","article-title":"Learning to optimize tensor programs","author":"chen","year":"2018","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref58","first-page":"863","article-title":"Ansor: Generating {High-Performance}tensor programs for deep learning","author":"zheng","year":"2020","journal-title":"Proc 14th USENIX Symp Operating Syst Des Implementation"},{"article-title":"Huggingface's transformers: State-of-the-art natural language processing","year":"2019","author":"wolf","key":"ref53"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/GreenCom-CPSCom.2010.102"},{"article-title":"Layer normalization","year":"2016","author":"ba","key":"ref11"},{"key":"ref55","first-page":"951","article-title":"DeepCPU: Serving RNN-based deep learning models 10x faster","author":"zhang","year":"2018","journal-title":"Proc USENIX Annu Tech Conf"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1145\/2858788.2688521"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1109\/EMC2-NIPS53020.2019.00016"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"journal-title":"Intel avx-512 instructions and their use in the implementation of math functions","year":"2015","author":"cornea","key":"ref16"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1007\/BF01386390"},{"article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","year":"2018","author":"devlin","key":"ref18"},{"key":"ref51","first-page":"5998","article-title":"Attention is all you need","author":"vaswani","year":"2017","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref50","first-page":"10347","article-title":"Training data-efficient image transformers & distillation through attention","author":"touvron","year":"2021","journal-title":"Proc Int Conf Mach Learn"},{"article-title":"FastSpeech 2: Fast and high-quality end-to-end text to speech","year":"2020","author":"ren","key":"ref46"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3406703"},{"article-title":"An image is worth 16x16 words, what is a video worth?","year":"2021","author":"sharir","key":"ref48"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00353"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1109\/CGO.2019.8661176"},{"key":"ref41","first-page":"8026","article-title":"PyTorch: An imperative style, high-performance deep learning library","author":"paszke","year":"2019","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D16-1264"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/ASAP.2019.00-42"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1093\/comjnl\/40.6.340"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1016\/j.suscom.2022.100683"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1145\/3306346.3322967"},{"key":"ref9","doi-asserted-by":"crossref","DOI":"10.1093\/oso\/9780198565581.001.0001","author":"andrews","year":"1998","journal-title":"Special Functions of Mathematics for Engineers"},{"year":"2022","key":"ref4","article-title":"microsoft\/onnxruntime: ONNX runtime: Cross-platform, high performance ML inferencing and training accelerator"},{"year":"2021","key":"ref3","article-title":"onnxruntime\/onnxruntime\/core\/optimizer at main microsoft\/onnxruntime"},{"year":"2022","key":"ref6","article-title":"onnxruntime\/onnxruntime\/core\/mlas at master microsoft\/onnxruntime"},{"year":"2022","key":"ref5","article-title":"ONNX\/ONNX: Open standard for machine learning interoperability"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1145\/2786763.2694364"},{"article-title":"Roberta: A robustly optimized bert pretraining approach","year":"2019","author":"liu","key":"ref35"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1007\/s10586-018-2810-y"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2021.3071762"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.sustainlp-1.20"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1145\/3293320.3293334"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v33i01.33016706"},{"article-title":"Reformer: The efficient transformer","year":"2020","author":"kitaev","key":"ref32"},{"year":"2021","key":"ref2","article-title":"onnxruntime\/onnxruntime\/contrib_ops\/cpu\/bert at main microsoft\/onnxruntime"},{"year":"2017","key":"ref1","article-title":"A more accurate, performance-competitive implementation of expf() - cuda \/ cuda programming and performance - nvidia developer forums"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1145\/2897824.2925952"},{"article-title":"Transformers with convolutional context for ASR","year":"2019","author":"mohamed","key":"ref38"},{"article-title":"Gaussian error linear units (GELUs)","year":"2016","author":"hendrycks","key":"ref24"},{"key":"ref23","first-page":"281","article-title":"Collective loop fusion for array contraction","author":"gao","year":"1992","journal-title":"Proc Int Workshop Lang Compilers Parallel Comput"},{"key":"ref26","first-page":"711","article-title":"Data movement is all you need: A case study on optimizing transformers","author":"ivanov","year":"2021","journal-title":"Proc Mach Learn Syst Conf"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1145\/369028.369096"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2018.8462506"},{"article-title":"Memory-aware fusing and tiling of neural networks for accelerated edge inference","year":"2021","author":"farley","key":"ref22"},{"article-title":"An image is worth 16x16 words: Transformers for image recognition at scale","year":"2020","author":"dosovitskiy","key":"ref21"},{"key":"ref28","first-page":"301","article-title":"Maximizing loop parallelism and improving data locality via loop fusion and distribution","author":"kennedy","year":"1993","journal-title":"Proc Int Workshop Lang Compilers Parallel Comput"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1007\/s11227-015-1613-7"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1007\/s00607-021-00958-2"}],"container-title":["IEEE Transactions on Parallel and Distributed Systems"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/71\/10123122\/10107474.pdf?arnumber=10107474","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,12,11]],"date-time":"2023-12-11T10:29:06Z","timestamp":1702290546000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10107474\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,6]]},"references-count":58,"journal-issue":{"issue":"6"},"URL":"https:\/\/doi.org\/10.1109\/tpds.2023.3269530","relation":{},"ISSN":["1045-9219","1558-2183","2161-9883"],"issn-type":[{"type":"print","value":"1045-9219"},{"type":"electronic","value":"1558-2183"},{"type":"electronic","value":"2161-9883"}],"subject":[],"published":{"date-parts":[[2023,6]]}}}