{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,21]],"date-time":"2026-01-21T07:42:48Z","timestamp":1768981368832,"version":"3.49.0"},"reference-count":45,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2022,1,1]],"date-time":"2022-01-01T00:00:00Z","timestamp":1640995200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2022,1,1]],"date-time":"2022-01-01T00:00:00Z","timestamp":1640995200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2022,1,1]],"date-time":"2022-01-01T00:00:00Z","timestamp":1640995200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Comput."],"published-print":{"date-parts":[[2022]]},"DOI":"10.1109\/tc.2022.3209920","type":"journal-article","created":{"date-parts":[[2022,9,28]],"date-time":"2022-09-28T20:01:43Z","timestamp":1664395303000},"page":"1-13","source":"Crossref","is-referenced-by-count":4,"title":["Architecting a Flash-Based Storage System for Low-Cost Inference of Extreme-Scale DNNs"],"prefix":"10.1109","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-0292-3322","authenticated-orcid":false,"given":"Yunho","family":"Jin","sequence":"first","affiliation":[{"name":"Harvard University, Cambridge, MA, USA"}]},{"given":"Shine","family":"Kim","sequence":"additional","affiliation":[{"name":"Solution Product & Development, Memory Business, Samsung Electronics, Hwaseong-si, Gyeonggi-do, Republic of Korea"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2669-6849","authenticated-orcid":false,"given":"Tae Jun","family":"Ham","sequence":"additional","affiliation":[{"name":"Seoul National University, Gwanak-gu, Seoul, Republic of Korea"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4266-4919","authenticated-orcid":false,"given":"Jae W.","family":"Lee","sequence":"additional","affiliation":[{"name":"Seoul National University, Gwanak-gu, Seoul, Republic of Korea"}]}],"member":"263","reference":[{"key":"ref39","article-title":"Radeon VII 16 GB HBM 2 memory cost around ${\\$}$$320","year":"0"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA45697.2020.00045"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/SC.2016.47"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/2829974"},{"key":"ref31","first-page":"403","article-title":"D2FQ: Device-direct fair queueing for NVMe SSDs","author":"woo","year":"2021","journal-title":"Proc 19th USENIX Conf File Storage Technol"},{"key":"ref30","article-title":"Onfi: Open nand flash interface specifiaction 5.0","year":"0"},{"key":"ref37","first-page":"49","article-title":"MQSim: A framework for enabling realistic studies of modern multi-queue SSD devices","author":"tavakkol","year":"2018","journal-title":"Proc 16th USENIX Conf File Storage Technol USENIX Assoc"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1145\/3352460.3358252"},{"key":"ref35","first-page":"813","article-title":"Practical erase suspension for modern low-latency ssds","author":"kim","year":"2019","journal-title":"Proc USENIX Annu Tech Conf"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1145\/3445814.3446733"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1145\/3458817.3476205"},{"key":"ref40","article-title":"Newegg.com","year":"0"},{"key":"ref11","first-page":"387","article-title":"FlashNeuron: SSD-enabled large-batch training of very deep neural networks","author":"bae","year":"2021","journal-title":"Proc 19th USENIX Conf File Storage Technol"},{"key":"ref12","article-title":"Distributed hierarchical GPU parameter server for massive scale deep learning ads systems","author":"zhao","year":"2020"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/DSN.2015.49"},{"key":"ref14","article-title":"An Image is Worth 16x16 Words: Transformers for Image Recognition At Scale","author":"dosovitskiy","year":"2020"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/3357384.3357895"},{"key":"ref16","article-title":"turing-NLGMicrosoft","year":"2020"},{"key":"ref17","article-title":"Z-SSD SamsungSZ985","year":"2018"},{"key":"ref18","first-page":"121","article-title":"Scalable parallel flash firmware for many-core architectures","author":"zhang","year":"2020","journal-title":"Proc 18th USENIX Conf File Storage Technol"},{"key":"ref19","article-title":"NVIDIA A100 GPU","year":"2020"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/ISSCC42613.2021.9366054"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W18-6301"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA51647.2021.00056"},{"key":"ref3","article-title":"Tensor2Tensor for neural machine translation","author":"vaswani","year":"2018"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/K19-1074"},{"key":"ref29","first-page":"338","article-title":"A flash memory controller for 15 $\\mu$?s ultra-low-latency SSD using high-speed 3D NAND flash with 3 $\\mu$?s read time","author":"cheong","year":"2018","journal-title":"Proc IEEE Int Solid-State Circuits Conf"},{"key":"ref5","article-title":"Language models are unsupervised multitask learners","author":"radford","year":"2019","journal-title":"OpenAIRE blog"},{"key":"ref8","first-page":"6000","article-title":"Attention is all you need","author":"vaswani","year":"2017","journal-title":"Proc 30th Int Conf Neural Inf Process Syst"},{"key":"ref7","article-title":"Neural text summarization: A critical evaluation","author":"kry?ci?ski","year":"2019"},{"key":"ref2","article-title":"Language models are few-shot learners","author":"brown","year":"2020"},{"key":"ref9","first-page":"371","article-title":"Behemoth: A flash-centric training accelerator for extreme-scale DNNs","author":"kim","year":"2021","journal-title":"Proc 19th USENIX Conf File Storage Technol Assoc"},{"key":"ref1","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","author":"devlin","year":"2018"},{"key":"ref20","article-title":"PyTorch","year":"2021"},{"key":"ref45","first-page":"147","article-title":"Hardware\/Software co-programmable framework for computational SSDs to accelerate deep learning service on large-scale graphs","author":"kwon","year":"2022","journal-title":"Proc 20th USENIX Conf File Storage Technol"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2015.2504868"},{"key":"ref21","article-title":"Exploring the limits of transfer learning with a unified text-to-text Transformer","author":"raffel","year":"2019"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1145\/3007787.3001154"},{"key":"ref24","article-title":"Tensorflow: Large-scale machine learning on heterogeneous distributed systems","author":"abadi","year":"2016"},{"key":"ref41","article-title":"SSD price: Samsung SSD 970 EVO","year":"0"},{"key":"ref23","first-page":"281","article-title":"Design tradeoffs for SSD reliability","author":"kim","year":"2019","journal-title":"Proc 17th USENIX Conf File Storage Technol"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1145\/3445814.3446763"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1145\/3007787.3001177"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1145\/3352460.3358320"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1145\/3140659.3080246"}],"container-title":["IEEE Transactions on Computers"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/12\/4358213\/09904910.pdf?arnumber=9904910","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,11,17]],"date-time":"2022-11-17T16:33:04Z","timestamp":1668702784000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9904910\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022]]},"references-count":45,"URL":"https:\/\/doi.org\/10.1109\/tc.2022.3209920","relation":{},"ISSN":["0018-9340","1557-9956","2326-3814"],"issn-type":[{"value":"0018-9340","type":"print"},{"value":"1557-9956","type":"electronic"},{"value":"2326-3814","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022]]}}}