{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,6,18]],"date-time":"2025-06-18T04:29:58Z","timestamp":1750220998327,"version":"3.41.0"},"publisher-location":"New York, NY, USA","reference-count":22,"publisher":"ACM","license":[{"start":{"date-parts":[[2019,7,28]],"date-time":"2019-07-28T00:00:00Z","timestamp":1564272000000},"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":[],"published-print":{"date-parts":[[2019,7,28]]},"DOI":"10.1145\/3332186.3333047","type":"proceedings-article","created":{"date-parts":[[2019,7,30]],"date-time":"2019-07-30T19:55:41Z","timestamp":1564516541000},"page":"1-6","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":6,"title":["MagmaDNN"],"prefix":"10.1145","author":[{"given":"Daniel","family":"Nichols","sequence":"first","affiliation":[{"name":"University of Tennessee, Knoxville"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Kwai","family":"Wong","sequence":"additional","affiliation":[{"name":"University of Tennessee, Knoxville"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Stan","family":"Tomov","sequence":"additional","affiliation":[{"name":"University of Tennessee, Knoxville"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Lucien","family":"Ng","sequence":"additional","affiliation":[{"name":"The Chinese University of Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Sihan","family":"Chen","sequence":"additional","affiliation":[{"name":"The Chinese University of Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Alex","family":"Gessinger","sequence":"additional","affiliation":[{"name":"Slippery Rock University"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"320","published-online":{"date-parts":[[2019,7,28]]},"reference":[{"key":"e_1_3_2_1_1_1","unstructured":"1998. MNIST. http:\/\/yann.lecun.com\/exdb\/mnist\/. http:\/\/yann.lecun.com\/exdb\/mnist\/ {Online; accessed 11-March-2019}.  1998. MNIST. http:\/\/yann.lecun.com\/exdb\/mnist\/. http:\/\/yann.lecun.com\/exdb\/mnist\/ {Online; accessed 11-March-2019}."},{"key":"e_1_3_2_1_2_1","unstructured":"2011. ImageNet. http:\/\/imagenet.org\/index. http:\/\/image-net.org\/index {Online; accessed 11-March-2019}.  2011. ImageNet. http:\/\/imagenet.org\/index. http:\/\/image-net.org\/index {Online; accessed 11-March-2019}."},{"volume-title":"Chris Olah","author":"Abadi Mart\u00edn","key":"e_1_3_2_1_3_1","unstructured":"Mart\u00edn Abadi , Ashish Agarwal , Paul Barham , Eugene Brevdo , Zhifeng Chen , Craig Citro , Gregory S. Corrado , Andy Davis , Jeffrey Dean , Matthieu Devin , Sanjay Ghemawat , Ian J. Goodfellow , Andrew Harp , Geoffrey Irving , Michael Isard , Yangqing Jia , Rafal J\u00f3zefowicz , Lukasz Kaiser , Manjunath Kudlur , Josh Levenberg , Dan Man\u00e9 , Rajat Monga , Sherry Moore , Derek Gordon Murray , Chris Olah , Mike Schuster , Jonathon Shlens , Benoit Steiner, Ilya Sutskever, Kunal Talwar, Paul A. Tucker, Vincent Vanhoucke, Vijay Vasudevan, Fernanda B. Vi\u00e9gas, Oriol Vinyals, Pete Warden, Martin Wattenberg, Martin Wicke, Yuan Yu, and Xiaoqiang Zheng. 2016. TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems. CoRR abs\/1603.04467 (2016). arXiv:1603.04467 http:\/\/arxiv.org\/abs\/1603.04467 Mart\u00edn Abadi, Ashish Agarwal, Paul Barham, Eugene Brevdo, Zhifeng Chen, Craig Citro, Gregory S. Corrado, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Ian J. Goodfellow, Andrew Harp, Geoffrey Irving, Michael Isard, Yangqing Jia, Rafal J\u00f3zefowicz, Lukasz Kaiser, Manjunath Kudlur, Josh Levenberg, Dan Man\u00e9, Rajat Monga, Sherry Moore, Derek Gordon Murray, Chris Olah, Mike Schuster, Jonathon Shlens, Benoit Steiner, Ilya Sutskever, Kunal Talwar, Paul A. Tucker, Vincent Vanhoucke, Vijay Vasudevan, Fernanda B. Vi\u00e9gas, Oriol Vinyals, Pete Warden, Martin Wattenberg, Martin Wicke, Yuan Yu, and Xiaoqiang Zheng. 2016. TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems. CoRR abs\/1603.04467 (2016). arXiv:1603.04467 http:\/\/arxiv.org\/abs\/1603.04467"},{"key":"e_1_3_2_1_4_1","volume-title":"Demystifying Parallel and Distributed Deep Learning: An In-Depth Concurrency Analysis. CoRR abs\/1802.09941","author":"Ben-Nun Tal","year":"2018","unstructured":"Tal Ben-Nun and Torsten Hoefler . 2018. Demystifying Parallel and Distributed Deep Learning: An In-Depth Concurrency Analysis. CoRR abs\/1802.09941 ( 2018 ). arXiv:1802.09941 http:\/\/arxiv.org\/abs\/1802.09941 Tal Ben-Nun and Torsten Hoefler. 2018. Demystifying Parallel and Distributed Deep Learning: An In-Depth Concurrency Analysis. CoRR abs\/1802.09941 (2018). arXiv:1802.09941 http:\/\/arxiv.org\/abs\/1802.09941"},{"key":"e_1_3_2_1_5_1","volume-title":"Davide Del Testa","author":"Bojarski Mariusz","year":"2016","unstructured":"Mariusz Bojarski , Davide Del Testa , Daniel Dworakowski, Bernhard Firner , Beat Flepp, Prasoon Goyal, Lawrence D. Jackel, Mathew Monfort, Urs Muller, Jiakai Zhang, and et al. 2016 . End to End Learning for Self-Driving Cars . arXiv:1604.07316 {cs} (Apr 2016). http:\/\/arxiv.org\/abs\/1604.07316 arXiv: 1604.07316. Mariusz Bojarski, Davide Del Testa, Daniel Dworakowski, Bernhard Firner, Beat Flepp, Prasoon Goyal, Lawrence D. Jackel, Mathew Monfort, Urs Muller, Jiakai Zhang, and et al. 2016. End to End Learning for Self-Driving Cars. arXiv:1604.07316 {cs} (Apr 2016). http:\/\/arxiv.org\/abs\/1604.07316 arXiv: 1604.07316."},{"key":"e_1_3_2_1_6_1","volume-title":"Revisiting Distributed Synchronous SGD. CoRR abs\/1604.00981","author":"Chen Jianmin","year":"2016","unstructured":"Jianmin Chen , Rajat Monga , Samy Bengio , and Rafal J\u00f3zefowicz . 2016. Revisiting Distributed Synchronous SGD. CoRR abs\/1604.00981 ( 2016 ). arXiv:1604.00981 http:\/\/arxiv.org\/abs\/1604.00981 Jianmin Chen, Rajat Monga, Samy Bengio, and Rafal J\u00f3zefowicz. 2016. Revisiting Distributed Synchronous SGD. CoRR abs\/1604.00981 (2016). arXiv:1604.00981 http:\/\/arxiv.org\/abs\/1604.00981"},{"key":"e_1_3_2_1_7_1","volume-title":"cuDNN: Efficient Primitives for Deep Learning. CoRR abs\/1410.0759","author":"Chetlur Sharan","year":"2014","unstructured":"Sharan Chetlur , Cliff Woolley , Philippe Vandermersch , Jonathan Cohen , John Tran , Bryan Catanzaro , and Evan Shelhamer . 2014. cuDNN: Efficient Primitives for Deep Learning. CoRR abs\/1410.0759 ( 2014 ). arXiv:1410.0759 http:\/\/arxiv.org\/abs\/1410.0759 Sharan Chetlur, Cliff Woolley, Philippe Vandermersch, Jonathan Cohen, John Tran, Bryan Catanzaro, and Evan Shelhamer. 2014. cuDNN: Efficient Primitives for Deep Learning. CoRR abs\/1410.0759 (2014). arXiv:1410.0759 http:\/\/arxiv.org\/abs\/1410.0759"},{"key":"e_1_3_2_1_8_1","unstructured":"Adam Coates Brody Huval Tao Wang David J Wu Andrew Y Ng and Bryan Catanzaro. {n. d.}. Deep learning with COTS HPC systems. ({n. d.}) 9.  Adam Coates Brody Huval Tao Wang David J Wu Andrew Y Ng and Bryan Catanzaro. {n. d.}. Deep learning with COTS HPC systems. ({n. d.}) 9."},{"volume-title":"Deep Learning","author":"Goodfellow Ian","key":"e_1_3_2_1_9_1","unstructured":"Ian Goodfellow , Yoshua Bengio , and Aaron Courville . 2016. Deep Learning . MIT Press . http:\/\/www.deeplearningbook.org. Ian Goodfellow, Yoshua Bengio, and Aaron Courville. 2016. Deep Learning. MIT Press. http:\/\/www.deeplearningbook.org."},{"key":"e_1_3_2_1_10_1","volume-title":"Large Minibatch SGD: Training ImageNet in 1 Hour. CoRR abs\/1706.02677","author":"Goyal Priya","year":"2017","unstructured":"Priya Goyal , Piotr Doll\u00e1r , Ross B. Girshick , Pieter Noordhuis , LukaszWesolowski, Aapo Kyrola , Andrew Tulloch , Yangqing Jia , and Kaiming He. 2017. Accurate , Large Minibatch SGD: Training ImageNet in 1 Hour. CoRR abs\/1706.02677 ( 2017 ). arXiv:1706.02677 http:\/\/arxiv.org\/abs\/1706.02677 Priya Goyal, Piotr Doll\u00e1r, Ross B. Girshick, Pieter Noordhuis, LukaszWesolowski, Aapo Kyrola, Andrew Tulloch, Yangqing Jia, and Kaiming He. 2017. Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour. CoRR abs\/1706.02677 (2017). arXiv:1706.02677 http:\/\/arxiv.org\/abs\/1706.02677"},{"key":"e_1_3_2_1_11_1","doi-asserted-by":"publisher","DOI":"10.1109\/MSP.2012.2205597"},{"key":"e_1_3_2_1_12_1","volume-title":"FireCaffe: near-linear acceleration of deep neural network training on compute clusters. CoRR abs\/1511.00175","author":"Iandola Forrest N.","year":"2015","unstructured":"Forrest N. Iandola , Khalid Ashraf , Matthew W. Moskewicz , and Kurt Keutzer . 2015. FireCaffe: near-linear acceleration of deep neural network training on compute clusters. CoRR abs\/1511.00175 ( 2015 ). arXiv:1511.00175 http:\/\/arxiv.org\/abs\/1511.00175 Forrest N. Iandola, Khalid Ashraf, Matthew W. Moskewicz, and Kurt Keutzer. 2015. FireCaffe: near-linear acceleration of deep neural network training on compute clusters. CoRR abs\/1511.00175 (2015). arXiv:1511.00175 http:\/\/arxiv.org\/abs\/1511.00175"},{"key":"e_1_3_2_1_13_1","doi-asserted-by":"publisher","DOI":"10.1145\/2647868.2654889"},{"volume-title":"ImageNet Classification with Deep Convolutional Neural Networks. Curran Associates","author":"Krizhevsky Alex","key":"e_1_3_2_1_14_1","unstructured":"Alex Krizhevsky , Ilya Sutskever , and Geoffrey E Hinton . 2012. ImageNet Classification with Deep Convolutional Neural Networks. Curran Associates , Inc ., 1097\u00e2\u0102\u015e1105. http:\/\/papers.nips.cc\/paper\/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. 2012. ImageNet Classification with Deep Convolutional Neural Networks. Curran Associates, Inc., 1097\u00e2\u0102\u015e1105. http:\/\/papers.nips.cc\/paper\/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf"},{"key":"e_1_3_2_1_15_1","volume-title":"Deep learning. Nature 521 (May","author":"LeCun Yann","year":"2015","unstructured":"Yann LeCun , Yoshua Bengio , and Geoffrey Hinton . 2015. Deep learning. Nature 521 (May 2015 ), 436. Citation Key : lecunDeepLearning2015a. Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. 2015. Deep learning. Nature 521 (May 2015), 436. Citation Key: lecunDeepLearning2015a."},{"key":"e_1_3_2_1_16_1","doi-asserted-by":"publisher","DOI":"10.1109\/5.726791"},{"key":"e_1_3_2_1_17_1","unstructured":"MagmaDNN 2019. MagmaDNN Site. https:\/\/bitbucket.org\/icl\/magmadnn.  MagmaDNN 2019. MagmaDNN Site. https:\/\/bitbucket.org\/icl\/magmadnn."},{"key":"e_1_3_2_1_18_1","doi-asserted-by":"publisher","DOI":"10.1038\/nature16961"},{"key":"e_1_3_2_1_19_1","volume-title":"Le","author":"Smith Samuel L.","year":"2017","unstructured":"Samuel L. Smith , Pieter-Jan Kindermans , and Quoc V . Le . 2017 . Don't Decay the Learning Rate, Increase the Batch Size. CoRR abs\/1711.00489 (2017). arXiv:1711.00489 http:\/\/arxiv.org\/abs\/1711.00489 Samuel L. Smith, Pieter-Jan Kindermans, and Quoc V. Le. 2017. Don't Decay the Learning Rate, Increase the Batch Size. CoRR abs\/1711.00489 (2017). arXiv:1711.00489 http:\/\/arxiv.org\/abs\/1711.00489"},{"key":"e_1_3_2_1_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.parco.2009.12.005"},{"key":"e_1_3_2_1_21_1","volume-title":"Scaling SGD Batch Size to 32K for ImageNet Training. CoRR abs\/1708.03888","author":"You Yang","year":"2017","unstructured":"Yang You , Igor Gitman , and Boris Ginsburg . 2017. Scaling SGD Batch Size to 32K for ImageNet Training. CoRR abs\/1708.03888 ( 2017 ). arXiv:1708.03888 http:\/\/arxiv.org\/abs\/1708.03888 Yang You, Igor Gitman, and Boris Ginsburg. 2017. Scaling SGD Batch Size to 32K for ImageNet Training. CoRR abs\/1708.03888 (2017). arXiv:1708.03888 http:\/\/arxiv.org\/abs\/1708.03888"},{"key":"e_1_3_2_1_22_1","volume-title":"100-epoch ImageNet Training with AlexNet in 24 Minutes. CoRR abs\/1709.05011","author":"You Yang","year":"2017","unstructured":"Yang You , Zhao Zhang , Cho-Jui Hsieh , and James Demmel . 2017. 100-epoch ImageNet Training with AlexNet in 24 Minutes. CoRR abs\/1709.05011 ( 2017 ). arXiv:1709.05011 http:\/\/arxiv.org\/abs\/1709.05011 Yang You, Zhao Zhang, Cho-Jui Hsieh, and James Demmel. 2017. 100-epoch ImageNet Training with AlexNet in 24 Minutes. CoRR abs\/1709.05011 (2017). arXiv:1709.05011 http:\/\/arxiv.org\/abs\/1709.05011"}],"event":{"name":"PEARC '19: Practice and Experience in Advanced Research Computing","acronym":"PEARC '19","location":"Chicago IL USA"},"container-title":["Proceedings of the Practice and Experience in Advanced Research Computing on Rise of the Machines (learning)"],"original-title":[],"link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3332186.3333047","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3332186.3333047","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,18]],"date-time":"2025-06-18T00:25:38Z","timestamp":1750206338000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3332186.3333047"}},"subtitle":["Accelerated Deep Learning Using MAGMA"],"short-title":[],"issued":{"date-parts":[[2019,7,28]]},"references-count":22,"alternative-id":["10.1145\/3332186.3333047","10.1145\/3332186"],"URL":"https:\/\/doi.org\/10.1145\/3332186.3333047","relation":{},"subject":[],"published":{"date-parts":[[2019,7,28]]},"assertion":[{"value":"2019-07-28","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}