each Chrome performance

[ghost]

I get horrible performance from $.each() in Chrome.

Can I request that it no longer be used for new loops?

If's it's alright, I'll push some conversions for the old ones.

Thanks.

[Mottie]

Hi @jimbobcoder!

Sure! I would appreciate any help :)

[Mercuenomia]

Count me in.

Side note: thank you very much for this library. Lifesaver.

Let me know if you need help animating Pager.

[Mottie]

I was looking through the code, and really the only place it looks like I use $.each() is for small arrays, like processing selected widgets and updating headers. Usually arrays the same size as the number of columns in the table. The only code that looks like it might show a noticeable benefit would be the pager code - creating options for the page selector & calculating the number of filtered rows.

I'm all for performance enhancement, but I don't see where in the current code there would be a major issue.

Don't get me wrong, I still want to change things over to plain javascript. I was looking at this jspref and found it surprising that $.each() is faster than for...in and array.forEach, but still much slower than a for loop.

[ghost]

Yes, it looks like it could be finished relatively quickly since they're compact tight loops.

For in must execute a function or Chrome will not optimize. I can't speak on foreach.

Sampling the length once and looping via index is the fastest, I've found. $.each previously used just about anything, so that could explain its middling type results. They may be optimizing their fors, hopefully.

The optimizer identifies trigger for me.

[Mottie]

I took care of the core plugin, filter widget and pager (addon & widget) so far.

I'll try to get to the other widgets soon.

[Mercuenomia]

Thanks! My tables are even faster now!

[ghost]

Wow, that was fast.

Would you like me to convert the eaches on the jQuery objects?

[Mottie]

From that jsPref test, it seems the jQuery each is faster with objects. I don't know how that is possible, but most of the remaining .each()'s are on objects.

[ghost]

For smaller sizes, it look like they perform equally.

For larger sizes, it looks like there's still a large percentage increase in performance converting to an index loop.

Which revision shows each as faster?

[Mottie]

The one I linked before...

What I was trying to say is that there are no large sizes. The largest object that is being run through the $.each() would match the number of columns in a table. So unless you have a table with 500 columns, I don't think you would notice a differece.

[ghost]

I just tested, and the compiler is still complaining that it cannot optimize the object eaches "ForInStatement is not fast case", so it looks like jQuery is using the old for in without wrapping the contents in a function.

Easy fix might be to wrap the contents of the object eaches in functions? I'll try and then get back to you as soon as I've tested.