Arun A. Balakrishnan, S. Kamal, C. SatheeshChandran
{"title":"基于模块化微服务的GPU利用率管理器","authors":"Arun A. Balakrishnan, S. Kamal, C. SatheeshChandran","doi":"10.46243/jst.2020.v5.i4.pp230-237","DOIUrl":null,"url":null,"abstract":":Graphics processing unit (GPU) is a computer programmable chip that could perform rapid\nmathematical operations that can be accelerated with massive parallelism. In the early days, central processing unit\n(CPU) was responsible for all computations irrespective of whether it is feasible for parallel computation. However,\nin recent years GPUs are increasingly used for massively parallel computing applications, such as training Deep\nNeural Networks. GPU’s performance monitoring plays a key role in this new era since GPUs serve an inevitable\nrole in increasing the speed of analysis of the developed system. GPU administration comes in picture to efficiently\nutilize the GPU when we deal with multiple workloads to run on the same hardware. In this study, various GPUparameters are monitored and help to keep them in safe levels and also to keep the improved performance of the\nsystem. This study,","PeriodicalId":23534,"journal":{"name":"Volume 5, Issue 4","volume":"83 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2020-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Modular Microservice based GPU Utilization Manager with\\nGunicorn\",\"authors\":\"Arun A. Balakrishnan, S. Kamal, C. SatheeshChandran\",\"doi\":\"10.46243/jst.2020.v5.i4.pp230-237\",\"DOIUrl\":null,\"url\":null,\"abstract\":\":Graphics processing unit (GPU) is a computer programmable chip that could perform rapid\\nmathematical operations that can be accelerated with massive parallelism. In the early days, central processing unit\\n(CPU) was responsible for all computations irrespective of whether it is feasible for parallel computation. However,\\nin recent years GPUs are increasingly used for massively parallel computing applications, such as training Deep\\nNeural Networks. GPU’s performance monitoring plays a key role in this new era since GPUs serve an inevitable\\nrole in increasing the speed of analysis of the developed system. GPU administration comes in picture to efficiently\\nutilize the GPU when we deal with multiple workloads to run on the same hardware. In this study, various GPUparameters are monitored and help to keep them in safe levels and also to keep the improved performance of the\\nsystem. This study,\",\"PeriodicalId\":23534,\"journal\":{\"name\":\"Volume 5, Issue 4\",\"volume\":\"83 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2020-07-30\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Volume 5, Issue 4\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.46243/jst.2020.v5.i4.pp230-237\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Volume 5, Issue 4","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.46243/jst.2020.v5.i4.pp230-237","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Modular Microservice based GPU Utilization Manager with
Gunicorn
:Graphics processing unit (GPU) is a computer programmable chip that could perform rapid
mathematical operations that can be accelerated with massive parallelism. In the early days, central processing unit
(CPU) was responsible for all computations irrespective of whether it is feasible for parallel computation. However,
in recent years GPUs are increasingly used for massively parallel computing applications, such as training Deep
Neural Networks. GPU’s performance monitoring plays a key role in this new era since GPUs serve an inevitable
role in increasing the speed of analysis of the developed system. GPU administration comes in picture to efficiently
utilize the GPU when we deal with multiple workloads to run on the same hardware. In this study, various GPUparameters are monitored and help to keep them in safe levels and also to keep the improved performance of the
system. This study,