E-Blah Speed as Compared to Other Forums:

As you know E-Blah has always been faster than most of the competion, so I did a little survey of the two and here's what I came up with.  These stats were taken September 2nd, 2005, 1:43pm.

This is all done before gZip compression has a time to work (as it'll use more CPU and Memory, although not much).  Both were done on a test system (mine) at 2.54GHz with 768MB ram.  The YaBB 2 forum was a clean install, as was E-Blah.  The only code added was the benchmark code.

First thing, I want to say that I used the latest source code (which is newer than P9.2). These are done WITHOUT modification!  The result is found by refreshing 10 times over every 1-2 seconds or so.  I was logged out on EACH board.

Here are the results:
Platinum 9.2+
Board Index: 0.07348 seconds load time (I would like to add this, as it's my test board with many boards and all; it's average is around 0.20?? or so) -- YaBB's default is two boards with a message posted
Message Index: 0.06784 seconds load time (note: I posted the EXACT message that was on the YaBB forum, just to make them both the same)
Message Display: 0.07285 seconds load time (also, same message as was on YaBB forum)

YaBB 2 RC3 (Latest as of Sept 2nd, 2005)
Board Index: 0.26272
Message Index: 0.26527
Message Display: 0.27866

Now ..... these are all, as always, unbiased numbers.  I didn't go and run CDex and rip a few CD's when testing YaBB and just let my CPU idle durring E-Blah's testing.  It was unbiased, and I think y'all know that.  I threw out a number which was 0.3000 from YaBB because I went to their site and then the forum and it went that high -- hence, unbiased.

If anyone wants to know the memory usage ... roughly 8.5MB's is used for YaBB's board index; while only 4.9 (or 5) MB is used for E-Blah's.  YaBB's Message Index uses about: 8.5MB for one message (displaying only one message); while E-Blah uses: 4.7 (or about 5) MB.

Finally the message display for YaBB is: 8.4 (or about 8.5) MB; while E-Blah is about 4.8 (or 5) MB.

Again, all biased numbers and how I got them was by looking at the Windows Task manager and letting perl sleep before the output was sent to the browsers.  I could do this with other forums, but this has taken quite a while.