• This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn more.

Comparing the speed of Xenforo to the big 4

Digital Doctor

Well-known member
#2
I dont't think that would be a valid extrapolation to how fast phpBB3 is compared to xenforo.
Pearl: Look at the Notice on xenforo.com: it says Welcome to XenFans.com !
 

CTXMedia

Formerly CyclingTribe
#6
A proper test would be to use the same data set on all four packages on the same hardware.

Even then, the server itself can be tweaked and optimised to suit the specific package you are using as all four packages will behave differently, cache differently, use different numbers of files, and have different lengths/numbers/frequency of queries.

Cheers,
Shaun :D
 

CTXMedia

Formerly CyclingTribe
#7
A point in case was my move from vB3.8.4 to IPB3.1

I made no changes to the server and after migration everything ran much, much slower (up to 7 seconds page loads at times!!) and after coming from a very fast and efficient vB3 install, this was a nightmare.

However, over the coming months I found that by removing one particular plugin it shaved 3 seconds from the page load time, I tweaked MySQL to make it more efficient (for the IPB setup I now had), I added more memory to the server, I added an Opcode cache, I tweak memcached, I changed some settings within IPB itself to remove more complex queries that didn't add much to the overall experience for users, I used my second domain to setup a self-CDN for static content, I moved some of the Java tags further down the HTML of the templates, I modified Apache to use expires headers for some static content, etc. etc.

A constant shaving of milliseconds here and there which resulted in a very fast board that made me and my members much happier. (Second and subsequent page loads are much quicker too - the above test doesn't take that into account!! ;) ).

Cheers,
Shaun :D
 

Xarcell

Well-known member
#8
I think a more valid test would be to put all softwares on the same hardware/location. Perhaps someone here with dedicated server and access to all four softwares can make a fresh install of each, then test. If possible, make a script that adds the same 1mil posts and 10k members to the db on each software and test again. Although, all that may just be too much work to be worth the test.

From my personal experience xF is by far faster, especially with more posts & members.

EDIT: I also think more software should be added to the test. Such as vanilla and SMF.
 

MGSteve

Well-known member
#9
Xenforo 1.1 seems to be introducing many new features which may improve page loading and speed improvements (I believe) so I thought it would be interesting to see how Xenforo compares to some other discussion boards on the net.

http://www.webpagetest.org/video/view.php?id=110921_f1cb6e1eaae5f753eb496000736b4a22b633548b

Not a shock to see who came in last place, but it was interesting to see that xenforo came in 3rd place.
Complete tosh I'm afraid. As Alex said, far too many variables to directly compare.

If they were all running off the same server with the same dataset, then fine. Otherwise, its pointless and worse, misleading.

For example, I suspect that server is in the US, hence why XenForo is slow compared to the others who are based in the US as the XF server is UK based.
 

Marc

Well-known member
#12
This is excellent. Please could you do a test of the Apples I have here against my Oranges to see which taste more like a Pear ... Cheers for your help

:p
 

Rigel Kentaurus

Well-known member
#14
A point in case was my move from vB3.8.4 to IPB3.1

I made no changes to the server and after migration everything ran much, much slower (up to 7 seconds page loads at times!!) and after coming from a very fast and efficient vB3 install, this was a nightmare.

However, over the coming months I found that by removing one particular plugin it shaved 3 seconds from the page load time, I tweaked MySQL to make it more efficient (for the IPB setup I now had), I added more memory to the server, I added an Opcode cache, I tweak memcached, I changed some settings within IPB itself to remove more complex queries that didn't add much to the overall experience for users, I used my second domain to setup a self-CDN for static content, I moved some of the Java tags further down the HTML of the templates, I modified Apache to use expires headers for some static content, etc. etc.

A constant shaving of milliseconds here and there which resulted in a very fast board that made me and my members much happier. (Second and subsequent page loads are much quicker too - the above test doesn't take that into account!! ;) ).

Cheers,
Shaun :D
If you had to do all of that, it sounds horrible
I'd rather have a solution that works well out of the box, having to spend that much team tweaking the solution is good for a hobby or weekend project if you are a programmer, not too much if you are actually trying to build a community
 

CTXMedia

Formerly CyclingTribe
#15
If you had to do all of that, it sounds horrible
I'd rather have a solution that works well out of the box, having to spend that much team tweaking the solution is good for a hobby or weekend project if you are a programmer, not too much if you are actually trying to build a community
Much as it was quite a bit of work - what I learned from doing all of that has been really helpful, and has allowed me to get more out of my server than I previously was.

Granted, it would have been good if it had been a like-for-like shoe-in, but IPB with all the bells and whistles is an altogether different beast to a basic vB3.8.4 install ... (y)

The plus side is that once we're on XF it should really rock!! :)

Cheers,
Shaun :D
 

Trekkan

Well-known member
#16
I think that type of troubleshooting/optimization goes for anything, such as your OS. Yeah, things might slow down after a while, figure it out, clean it up, disable the crap, etc. The reason most things work out of the box is that depending on the product either everything is on (for maximum compatibility) or everything is off (to limit functionality and increase compatibility due to a more complex nature of an application).

Either way, it requires someone to go through it and improve it for the individual needs. So as "horrible" as it sounds, it's not uncommon at all, for pretty much any piece of software.
 

DeltaHF

Well-known member
#18
Odd, I thought I read it was UK based? Mind you on this BT Infinity most sites are bloody fast anyway, its hard to tell sometimes what's 'local' and what's not.
The company is based in the UK, the server that runs this website is located in Texas, according to a lookup of the IP address. It's most likely on a dedicated server with SoftLayer.
 

Mikey

Well-known member
#20
Texas is where my server is located as well, pings from there to xenforo.com are extremely fast, for exactly the reasons Alex made early on in the thread.

Fact is, as long as these comparison websites are around, people will do tests.. It's silly, really. These results mean nothing, as has been rehashed in this thread - to be a fair test (as we were all taught in science) the SITUATION has to be the same, with ONE variable (the software).. in these tests it is not so.

It's essentially like me linking you guys this chart as if to prove how great PHPBB is. Ridiculous really.