graphics performance 2019a vs 2016b

2020 4 23
1 个回答
回答已采纳
更新时间：2020 4 28
4 次查看（30 天）

0 个投票

running on a core I7, 16 GB ram 64 bit laptop. I have matlab m-files making figure plots with a couple dozen line handles encompassing 100's of thousands of points. 2016b performs faster in plotting them, while 2019a seems to hesitate between each set of lines. m-file code is no different between the two versions. the plots make use of the 'UserData' and 'Tag' functionality to store data in the lines, for use in extracting the data via selection/callbacks. Any obvious reason why 2019a is slower than 2016b in plotting?

请先登录，再回答此问题。

 采纳的回答

Yair Altman
Yair Altman 2020-4-26

0 个投票

R2019a uses the integrated axes toolbar, which constantly monitors mouse movements in order to generate data tips on-the-fly. If you turn off this functionality you'll probably discover that interactivity becomes much "snappier".
See http://undocumentedmatlab.com/articles/improving-graphics-interactivity

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

Christopher
Christopher 2020-4-28
thanks Yair, this did help. I'd been told of this before, but never really ran into the issue, so had not made it part of regular axes calls. I will from now on. I use a panel and cover it with tabs, similar to your example, but only populate the first two tabs, but each with about 20 line handles, each of which can have hundreds of thousands of points. The temporal performance was further slowed by what profiler found as unncessary calls to 'figure' and drawnow while prcessing each of the two tabs. (likely from debugging earlier on). Removing those sped up part of the code, and setting the interactivity to empty set improved it further. Thanks again, appreciate your expertise.

请先登录，再进行评论。

更多回答（0 个）

类别

在 帮助中心 和 File Exchange 中查找有关 Graphics Performance 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by