add_widget : more and more slow

[zserge1]

Hi everybody !

I want use Gridster for my web site, but i need to add lot of widgets with "add_widget" command. I do a test and there is a problem with "add_widget" function : grid is more and more slow and there are memory leak.

You can see that in this vidéo : http://www.dailymotion.com/video/xznqxq_gridsterslow_tech#.UYjO_crmWUk

However, as you can see on the video, if i add widget at beginning (not with add_widget function) there is no problem. Can you help me ? Something wrong with my code ?

Thanks in advance !

[dustmoo]

Acknowledged.

It appears to be a problem with the .next_position function witch loops through all grid positions and sees if there is a free space.

As the number of widgets increase this process is slower and slower.

This needs to be refactored so that the last position is managed somehow. I'll see if I can work this into my merge and update. You may be able to bypass the problem by passing in an explicit row and column.

[berseckercgn]

Hi there,
I'm having the same issue. Has this been addressed in the fork you created?
Thanks!

[berseckercgn]

Hey Dustin,

I dug a little into the grister.js adding console logging. I found that Coords.set is getting slower and slower with every added widget. I explicitly passed correct row and column to the add_widget call without any effect on the performance.

Also I noticed that removing widgets using the API does not speed up the grid again. Adding and removing 10 widgets 10 times makes the grid as slow as adding 100 items. Looks to me like some object/list/array(?) is not properly cleared when removing widgets.

Hope this info helps to nail down the issue.

Regards, Stefan

[dustmoo]

Thank you, there are definitely some memory leaks.
​
I have picked things up again, I'll be pushing some updates next week.

On Thu, Jul 11, 2013 at 9:17 AM, berseckercgn notifications@github.com
wrote:

Hey Dustin,
I dug a little into the grister.js adding console logging. I found that Coords.set is getting slower and slower with every added widget. I explicitly passed correct row and column to the add_widget call without any effect on the performance.
Also I noticed that removing widgets using the API does not speed up the grid again. Adding and removing 10 widgets 10 times makes the grid as slow as adding 100 items. Looks to me like some object/list/array(?) is not properly cleared when removing widgets.
Hope this info helps to nail down the issue.

Regards, Stefan

Reply to this email directly or view it on GitHub:
#166 (comment)

[berseckercgn]

Dustin,

I'm new to JQuery, but will make use of gridster (or a customized fork) in a project of mine. So if I can be of any help in a particular area here, please feel free to ping me.

Thanks.

[berseckercgn]

Have build this simple thing that shows the issue https://github.com/berseckercgn/gridster-perf-test - includes customized gridster_dbg.js with coarse console logging per function.

[dustmoo]

Nice, I will look into it.

On Fri, Jul 12, 2013 at 3:19 AM, berseckercgn notifications@github.com
wrote:

Have build this simple thing that shows the issue https://github.com/berseckercgn/gridster-perf-test - includes customized gridster_dbg.js with coarse console logging per function.

Reply to this email directly or view it on GitHub:
#166 (comment)

[fjbat]

Are there any updates on this?

[rrolla]

I also get headache with this 👎 i think, problem is in my code, but then i look here.

Is there some good solution?

[berseckercgn]

No solution yet, I doubt anyone is working on this currently.

[rehfeldchris]

workaround:

http://stackoverflow.com/questions/16418946/gridster-add-widget-is-slow

[willmlaughlin]

I had a massive slow down when I would unserialize from array and then attempt to add additional widgets. Don't know if anyone else had this problem, but here was my solution!

The 'rows' parameter in add_faux_rows was being interpreted as a string- this caused max_rows and actual_rows (used in a nested loop) to be incremented by a factor of 10 (i.e. 15 + '1' = '151', instead of 15 + 1 = 16). I solved this by adding a simple rows = parseInt(rows) at the top of the function. To be safe, I did the same for cols in the add_faux_cols function. Here's a fiddle with the updated functions:
http://jsfiddle.net/Les7j45L/

Hope this helps!

[jbraccini]

Neat work @willmlaughlin !!

[neofreko]

probably related to generate stylesheet function as well. I read on on other issue, porting the string concatenation into array join improves the performance significantly. I had slow performance issue when using 100 cols. However, using array join solved it.

[rochapablo]

Does anybody has any news?