Machine Learning Systems Are Stuck in a Rut. Barham, P. & Isard, M. In Proceedings of the Workshop on Hot Topics in Operating Systems, of HotOS '19, pages 177–183. ACM. Paper doi abstract bibtex In this paper we argue that systems for numerical computing are stuck in a local basin of performance and programmability. Systems researchers are doing an excellent job improving the performance of 5-year-old benchmarks, but gradually making it harder to explore innovative machine learning research ideas. We explain how the evolution of hardware accelerators favors compiler back ends that hyper-optimize large monolithic kernels, show how this reliance on high-performance but inflexible kernels reinforces the dominant style of programming model, and argue these programming abstractions lack expressiveness, maintainability, and modularity; all of which hinders research progress. We conclude by noting promising directions in the field, and advocate steps to advance progress towards high-performance general purpose numerical computing systems on modern accelerators.
@inproceedings{barhamMachineLearningSystems2019,
location = {{New York, NY, USA}},
title = {Machine {{Learning Systems Are Stuck}} in a {{Rut}}},
isbn = {978-1-4503-6727-1},
url = {http://doi.acm.org/10.1145/3317550.3321441},
doi = {10.1145/3317550.3321441},
abstract = {In this paper we argue that systems for numerical computing are stuck in a local basin of performance and programmability. Systems researchers are doing an excellent job improving the performance of 5-year-old benchmarks, but gradually making it harder to explore innovative machine learning research ideas. We explain how the evolution of hardware accelerators favors compiler back ends that hyper-optimize large monolithic kernels, show how this reliance on high-performance but inflexible kernels reinforces the dominant style of programming model, and argue these programming abstractions lack expressiveness, maintainability, and modularity; all of which hinders research progress. We conclude by noting promising directions in the field, and advocate steps to advance progress towards high-performance general purpose numerical computing systems on modern accelerators.},
booktitle = {Proceedings of the {{Workshop}} on {{Hot Topics}} in {{Operating Systems}}},
series = {{{HotOS}} '19},
publisher = {{ACM}},
urldate = {2019-06-07},
date = {2019},
pages = {177--183},
author = {Barham, Paul and Isard, Michael},
file = {/home/dimitri/Nextcloud/Zotero/storage/KK5T5X4X/Barham and Isard - 2019 - Machine Learning Systems Are Stuck in a Rut.pdf},
venue = {Bertinoro, Italy}
}
Downloads: 0
{"_id":"wvSLgNJR4TooQLqtR","bibbaseid":"barham-isard-machinelearningsystemsarestuckinarut","authorIDs":[],"author_short":["Barham, P.","Isard, M."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","location":"New York, NY, USA","title":"Machine Learning Systems Are Stuck in a Rut","isbn":"978-1-4503-6727-1","url":"http://doi.acm.org/10.1145/3317550.3321441","doi":"10.1145/3317550.3321441","abstract":"In this paper we argue that systems for numerical computing are stuck in a local basin of performance and programmability. Systems researchers are doing an excellent job improving the performance of 5-year-old benchmarks, but gradually making it harder to explore innovative machine learning research ideas. We explain how the evolution of hardware accelerators favors compiler back ends that hyper-optimize large monolithic kernels, show how this reliance on high-performance but inflexible kernels reinforces the dominant style of programming model, and argue these programming abstractions lack expressiveness, maintainability, and modularity; all of which hinders research progress. We conclude by noting promising directions in the field, and advocate steps to advance progress towards high-performance general purpose numerical computing systems on modern accelerators.","booktitle":"Proceedings of the Workshop on Hot Topics in Operating Systems","series":"HotOS '19","publisher":"ACM","urldate":"2019-06-07","date":"2019","pages":"177–183","author":[{"propositions":[],"lastnames":["Barham"],"firstnames":["Paul"],"suffixes":[]},{"propositions":[],"lastnames":["Isard"],"firstnames":["Michael"],"suffixes":[]}],"file":"/home/dimitri/Nextcloud/Zotero/storage/KK5T5X4X/Barham and Isard - 2019 - Machine Learning Systems Are Stuck in a Rut.pdf","venue":"Bertinoro, Italy","bibtex":"@inproceedings{barhamMachineLearningSystems2019,\n location = {{New York, NY, USA}},\n title = {Machine {{Learning Systems Are Stuck}} in a {{Rut}}},\n isbn = {978-1-4503-6727-1},\n url = {http://doi.acm.org/10.1145/3317550.3321441},\n doi = {10.1145/3317550.3321441},\n abstract = {In this paper we argue that systems for numerical computing are stuck in a local basin of performance and programmability. Systems researchers are doing an excellent job improving the performance of 5-year-old benchmarks, but gradually making it harder to explore innovative machine learning research ideas. We explain how the evolution of hardware accelerators favors compiler back ends that hyper-optimize large monolithic kernels, show how this reliance on high-performance but inflexible kernels reinforces the dominant style of programming model, and argue these programming abstractions lack expressiveness, maintainability, and modularity; all of which hinders research progress. We conclude by noting promising directions in the field, and advocate steps to advance progress towards high-performance general purpose numerical computing systems on modern accelerators.},\n booktitle = {Proceedings of the {{Workshop}} on {{Hot Topics}} in {{Operating Systems}}},\n series = {{{HotOS}} '19},\n publisher = {{ACM}},\n urldate = {2019-06-07},\n date = {2019},\n pages = {177--183},\n author = {Barham, Paul and Isard, Michael},\n file = {/home/dimitri/Nextcloud/Zotero/storage/KK5T5X4X/Barham and Isard - 2019 - Machine Learning Systems Are Stuck in a Rut.pdf},\n venue = {Bertinoro, Italy}\n}\n\n","author_short":["Barham, P.","Isard, M."],"key":"barhamMachineLearningSystems2019","id":"barhamMachineLearningSystems2019","bibbaseid":"barham-isard-machinelearningsystemsarestuckinarut","role":"author","urls":{"Paper":"http://doi.acm.org/10.1145/3317550.3321441"},"downloads":0},"bibtype":"inproceedings","biburl":"https://raw.githubusercontent.com/dlozeve/newblog/master/bib/all.bib","creationDate":"2020-01-08T20:39:39.374Z","downloads":0,"keywords":[],"search_terms":["machine","learning","systems","stuck","rut","barham","isard"],"title":"Machine Learning Systems Are Stuck in a Rut","year":null,"dataSources":["3XqdvqRE7zuX4cm8m"]}