vB Optimise Pro compile time stats

ZeroHour · 4th January 2012, 14:13

Hi
Just wondering if there were any examples of compile time (using the debugger numbers) to show how much vB Optimise Pro improves things on a default install?
Guest cache viewing page compile stats too would be very useful.
I am curios about the product but I would like some numbers on a default install to compare with.

Cheers

0 Comment(s)

Belazor · 4th January 2012, 14:15

That's not possible I'm afraid, as saving 1 query on one server reduces loading times by a different number than it does on another server.

This is the very same reason why you should never ever ask a developer "how much load does this mod add?" - there is no answer.

Sorry

0 Comment(s)

Decado · 4th January 2012, 14:50

You can get some stats for real-world performance for the mod from here: vB Optimise - Resource Statistics

0 Comment(s)

ZeroHour · 4th January 2012, 15:23

Originally Posted by Belazor

That's not possible I'm afraid, as saving 1 query on one server reduces loading times by a different number than it does on another server.

This is the very same reason why you should never ever ask a developer "how much load does this mod add?" - there is no answer.

Sorry

Yeh I know what you mean, its more just that its a lot of cash to shell out to find the compile time has not reduced. Saving queries is good but its just a factor in compile time tbh and is the lesser part of several types of pages which could mean I get minimal improvements.

Does guest cache show a big boost over page loading without? I presume it just uses the cacher installed (in our case xcache)?
I have been testing varnish/nginx cache but the cookies can be a problem.
Even a ball park % reduction of compile time before guest cache to with it enabled. I am testing the lite version but the guest cache is probably the best part of the product for us.

0 Comment(s)

Belazor · 4th January 2012, 17:13

PHP scripts don't compile, you mean page load time

We ran a couple of webpage tests:
vBO On ("Repeat View" hitting guest cache): WebPagetest Test Result - London : demo.dragonbyte-tech.com/ - 01/04/12 17:05:28
vBO Disabled: WebPagetest Test Result - London : demo.dragonbyte-tech.com/ - 01/04/12 17:10:58

It's pretty easy drawing a conclusion from that

0 Comment(s)

ZeroHour · 4th January 2012, 18:39

Originally Posted by Belazor

PHP scripts don't compile, you mean page load time

Hehe yeh just habbit calling all processing compiling, time to compile some tea

Originally Posted by Belazor

We ran a couple of webpage tests:
vBO On ("Repeat View" hitting guest cache): WebPagetest Test Result - London : demo.dragonbyte-tech.com/ - 01/04/12 17:05:28
vBO Disabled: WebPagetest Test Result - London : demo.dragonbyte-tech.com/ - 01/04/12 17:10:58

It's pretty easy drawing a conclusion from that

Thanks for those, that does help a lot.
The tool I was thinking of to show load time was the debugging stats at the bottom of pages when you enable vbulletin debug mode as that gives a good number for page generation.

0 Comment(s)

Belazor · 4th January 2012, 19:16

Those stats at the bottom are useless IMO, the actual rendering time as displayed in those test results matter more. Generating the page in 0.0001 seconds and having users spend 5 seconds downloading poorly compressed images... yeah xD

0 Comment(s)

ZeroHour · 4th January 2012, 21:57

Originally Posted by Belazor

Those stats at the bottom are useless IMO, the actual rendering time as displayed in those test results matter more. Generating the page in 0.0001 seconds and having users spend 5 seconds downloading poorly compressed images... yeah xD

In a way they are but the guest cache feature for example is one area where those stats are the relevant ones as it shows the difference in page generation which is what the cache helps with most.
Other elements loading/rendering too are very important but I dont think page generation time/time to first byte should be ignored especially as saying page generation from cache is a lot faster then without would only be seen as a benefit.

0 Comment(s)