{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,30]],"date-time":"2026-01-30T01:50:27Z","timestamp":1769737827024,"version":"3.49.0"},"reference-count":22,"publisher":"Association for Computing Machinery (ACM)","issue":"5s","license":[{"start":{"date-parts":[[2017,9,27]],"date-time":"2017-09-27T00:00:00Z","timestamp":1506470400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Embed. Comput. Syst."],"published-print":{"date-parts":[[2017,10,31]]},"abstract":"<jats:p>Despite their remarkable performance in various machine intelligence tasks, the computational intensity of Convolutional Neural Networks (CNNs) has hindered their widespread utilization in resource-constrained embedded and IoT systems. To address this problem, we present a framework for synthesis of efficient CNN inference software targeting mobile SoC platforms. We argue that thread granularity can substantially impact the performance and energy dissipation of the synthesized inference software, and demonstrate that launching the maximum number of logical threads, often promoted as a guiding principle by GPGPU practitioners, does not result in an efficient implementation for mobile SoCs. We hypothesize that the runtime of a CNN layer on a particular SoC platform can be accurately estimated as a linear function of its computational complexity, which may seem counter-intuitive, as modern mobile SoCs utilize a plethora of heterogeneous architectural features and dynamic resource management policies. Consequently, we develop a principled approach and a data-driven analytical model to optimize granularity of threads during CNN software synthesis. Experimental results with several modern CNNs mapped to a commodity Android smartphone with a Snapdragon SoC show up to 2.37X speedup in application runtime, and up to 1.9X improvement in its energy dissipation compared to existing approaches.<\/jats:p>","DOI":"10.1145\/3126555","type":"journal-article","created":{"date-parts":[[2017,9,27]],"date-time":"2017-09-27T12:33:53Z","timestamp":1506515633000},"page":"1-19","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":31,"title":["Machine Intelligence on Resource-Constrained IoT Devices"],"prefix":"10.1145","volume":"16","author":[{"given":"Mohammad","family":"Motamedi","sequence":"first","affiliation":[{"name":"University of California, Davis, CA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Daniel","family":"Fong","sequence":"additional","affiliation":[{"name":"University of California, Davis, CA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Soheil","family":"Ghiasi","sequence":"additional","affiliation":[{"name":"University of California, Davis, CA"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"320","published-online":{"date-parts":[[2017,9,27]]},"reference":[{"key":"e_1_2_1_1_1","unstructured":"2010. Copyright office provides exemption to DMCA. (2010). https:\/\/www.copyright.gov\/1201\/.  2010. Copyright office provides exemption to DMCA. (2010). https:\/\/www.copyright.gov\/1201\/."},{"key":"e_1_2_1_2_1","doi-asserted-by":"publisher","DOI":"10.1145\/1816038.1815993"},{"key":"e_1_2_1_3_1","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO.2014.58"},{"key":"e_1_2_1_4_1","volume-title":"NIPS Workshop.","author":"Collobert Ronan","year":"2011"},{"key":"e_1_2_1_5_1","volume-title":"Hardware-oriented Approximation of Convolutional Neural Networks. arXiv preprint arXiv:1604.03168","author":"Gysel Philipp","year":"2016"},{"key":"e_1_2_1_6_1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_2_1_7_1","doi-asserted-by":"publisher","DOI":"10.1109\/SASP.2008.4570792"},{"key":"e_1_2_1_8_1","volume-title":"SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and 0.5 MB model size. arXiv preprint arXiv:1602.07360","author":"Iandola Forrest N.","year":"2016"},{"key":"e_1_2_1_9_1","doi-asserted-by":"publisher","DOI":"10.5555\/2133429.2133431"},{"key":"e_1_2_1_10_1","doi-asserted-by":"publisher","DOI":"10.1145\/2647868.2654889"},{"key":"e_1_2_1_11_1","volume-title":"Google Blog","author":"Jouppi N.","year":"2016"},{"key":"e_1_2_1_12_1","doi-asserted-by":"publisher","DOI":"10.1145\/3079856.3080246"},{"key":"e_1_2_1_13_1","volume-title":"Hinton","author":"Krizhevsky Alex","year":"2012"},{"key":"e_1_2_1_14_1","doi-asserted-by":"publisher","DOI":"10.1145\/2964284.2973801"},{"key":"e_1_2_1_15_1","doi-asserted-by":"publisher","DOI":"10.1145\/2628071.2628087"},{"key":"e_1_2_1_16_1","doi-asserted-by":"publisher","DOI":"10.1145\/2503210.2503268"},{"key":"e_1_2_1_17_1","volume-title":"Fast and Energy-Efficient CNN Inference on IoT Devices. arXiv preprint arXiv:1611.07151","author":"Motamedi Mohammad","year":"2016"},{"key":"e_1_2_1_18_1","volume-title":"Cappuccino: Efficient Inference Software Synthesis for Mobile System-on-Chips. arXiv preprint arXiv:1707.02647","author":"Motamedi Mohammad","year":"2017"},{"key":"e_1_2_1_19_1","doi-asserted-by":"publisher","DOI":"10.1109\/ASPDAC.2016.7428073"},{"key":"e_1_2_1_20_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-015-0816-y"},{"key":"e_1_2_1_21_1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2015.7298594"},{"key":"e_1_2_1_22_1","doi-asserted-by":"publisher","DOI":"10.1145\/2684746.2689060"}],"container-title":["ACM Transactions on Embedded Computing Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3126555","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3126555","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,18]],"date-time":"2025-06-18T19:05:02Z","timestamp":1750273502000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3126555"}},"subtitle":["The Case of Thread Granularity Optimization for CNN Inference"],"short-title":[],"issued":{"date-parts":[[2017,9,27]]},"references-count":22,"journal-issue":{"issue":"5s","published-print":{"date-parts":[[2017,10,31]]}},"alternative-id":["10.1145\/3126555"],"URL":"https:\/\/doi.org\/10.1145\/3126555","relation":{},"ISSN":["1539-9087","1558-3465"],"issn-type":[{"value":"1539-9087","type":"print"},{"value":"1558-3465","type":"electronic"}],"subject":[],"published":{"date-parts":[[2017,9,27]]},"assertion":[{"value":"2007-02-01","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2009-06-01","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2017-09-27","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}