Just saw this today. I use nginx as a front end proxy for static pages and image serving, so not sure how much this would help. I'd be curious if anyone else has tried it and their results ...

http://googlewebmastercentral.blogspot. ... aster.html

I installed it but didn't have time to play with changes. I'll give it a more fair shot in the future, but I noticed immediate problems with my Wordpress installations. I didn't see any apachebench increases. I'd like to hear from someone who's played with it more though!

Thanks for the input. I am with you right now -- I'll wait until the kinks get ironed out to mess with getting it up and running.

Don't use mod_pagespeed on your VPS now, it will slow down your server. I have to reboot my server 3 times in 1 day. The developer said he will fix the problem in next version. I thought the current version was tested.

Is this still true? I just added it and it seems to be working well.

I tested out mod_pagespeed on a development server a while ago, but ultimately decided not to use it in production.

I had hoped it would improve performance, by combining external Javascript and CSS files and reducing file requests. Sometimes though, I found it actually made results slightly worse. I am not sure why.

I do have to thank the mod_pagespeed team though because they have some really great tips for improving performance. Those ultimately led me to make my own open source Javascript/CSS combiner, which you can find on Github if you're interested: https://github.com/jeffzimmerlin/Javasc ... S-Combiner

It is hard to tell. I tested some sample pages and saw typical results like this:

Pagespeed on:

Load Time
First View
3.293s
Repeat View
1.748s


Pagespeed off:

Load Time
First View
3.361s
Repeat View
1.836s

That isn't much, though of course it does add up. But what processing time is it costing and is it worth it? That is hard to figure on a live server.

~

Yes, like you mentioned, the difference was nothing huge. Usually only a couple of hundred milliseconds here and there. Perhaps it was my naiveness, but I was surprised it wasn't always for the better (sometimes the results were actually worse). I suppose there are a lot of variables involved though.