20150728_bdti_adreno_white_paper_final英文资料

20150728_bdti_adreno_white_paper_final英文资料

ID:39078146

大小:482.84 KB

页数:9页

时间:2019-06-24

20150728_bdti_adreno_white_paper_final英文资料_第1页
20150728_bdti_adreno_white_paper_final英文资料_第2页
20150728_bdti_adreno_white_paper_final英文资料_第3页
20150728_bdti_adreno_white_paper_final英文资料_第4页
20150728_bdti_adreno_white_paper_final英文资料_第5页
资源描述:

《20150728_bdti_adreno_white_paper_final英文资料》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、AnIndependentEvaluationofImplementingComputerVisionFunctionswithOpenCLontheQualcommAdreno420BythestaffofBerkeleyDesignTechnology,Inc.July2015OVERVIEWComputervisionalgorithmsarebecomingincreasinglyimportantinmobile,embedded,andwearabledevicesandapplic

2、ations.Thesecompute-intensiveworkloadsarechallengingtoimplementwithgoodperformanceandpower-efficiency.Inmanyapplications,implementingcriticalportionsofcomputervisionworkloadsonageneral-purposegraphicsprocessingunit(GPU)isanattractivesolution.Qualcomm

3、enablesprogrammingoftheAdrenoGPUinitsSnapdragonapplicationprocessorsviatheopenstandardOpenCLlanguageandAPI.OpenCLsupportenablesprogrammerstooffloadcomputervisionalgorithmkernelstotheGPU,whichinturnprovidesspeedandpower-consumptionadvantagesoveraCPUim

4、plementation.BDTIdevelopedanAndroidapplicationdemonstratingcomputervisionfunctionalityutilizingtheAdreno420GPUinQualcomm’sSnapdragon805.TheBDTIapplicationcanruncompute-intensivecomputervisionfunctionsontheGPUortheCPU,enablingcomparisonsofGPUandCPUper

5、formanceandpower-efficiency.ThispaperdiscussestheBDTIapplication,implementationandoptimizationtechniquesusedinitsdevelopment,andthesubstantialbenefitsobservedwhenoffloadingcompute-intensivekernelstotheGPU.©2015BerkeleyDesignTechnology,Inc.Page1Conten

6、tstheimprovementsinperformanceandpowerconsumptionobtainedusingtheQualcomm1.Introduction.......................................................2AdrenoGPUinthisway.2.AlgorithmOverview.........................................2TheBDTIBackgroundBlurOpenCL

7、Android3.ImplementationOverview..............................4applicationwasdesignedtorunonthe4.AdrenoGPUOpenCL-AcceleratedSnapdragon805MDPtabletreferencedesignImplementation.................................................5fromIntrinsyc.Theapplicati

8、ondetectsthelargest5.ARMCPUNEON-Acceleratedforegroundobjectinthecamera’sview—usuallyImplementation.................................................8theuser—anddisplaysthevideowiththe6.BenefitsofOpenCLAccelerationontheforegroundobjectshownunaltered,an

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。