Script is slow, How can I make it faster?

[TimTimTimma]

As the title states, I have a script thats terribly slow on the initial visit, however after the sessions are set its fairly fast, or faster. But how can I set it up to run faster on the initial visit?

Below is the code.
I apologize for any indentation that is wild or not so good, I am using DreamWeaver to develop my scripts, and I cant find where to disable "Smart Tabs"

http://phpublisher.net/code/index.txt

[hitman6003]

Is there a session_start(); anywhere?

[TimTimTimma]

yes in the main script where its included, the code is as follows

if(!session_id()){
session_start();
}

[halojoy]

Hello

1. size is about 20.000 bytes
   can be cut down to 50% 10kB, by using short temp vars,
   instead of all them long: $SESSION['agagagaj'][ [$SESSION['some'] ]

How stupid halojoy would do it

<?php
$aids = $_SESSION[ 'article_ID' ];
$timg = $_SESSION[ 'Topic_Img' ];
$var1 = $_SESSION[ 'some var1' ];
$var2 = $_SESSION[ 'some var2' ];
// here begins testing 
if( $aids ){
get_a_cure_for_it();
}else{
if( $timg ) display_image( $timg );
else echo 'no img';
}
?>

2. to let php.exe call another program ( mysql(some function out side php.exe))
of course takes some extra time
So make as FEW calls as possible to mysql. Just one call to get a special data.
If it is used in several places, save it from FIRST call. In temporary short names variables
(I would of corse use FAST FLAT FILE Database storage, which is mastered
from inside php.exe - but that is only stupid HALOJOY way )

3. Can we do something to alter structure in IF ( conditions) ??
   I sometimes think about:
   Which cases (if-conditions) a more probable? - happens more often than other.
   If-conditions that are NOT so often happening,
   should only have to be tested for, if is such rare event

example: variable can be 0-9, but 90% of the time will be '9'
A)
switch (var)
case 1:
case 2
...
case 8:
case 9: do; break;

😎
switch (var)
case 9: do; break;
case 1
case 2
...
case 8

😎 will certainly improve speed, in 90% of all cases (when var=9)

#############################

These were some thought from me.
I would say, I would be able to cut down
your script AVERAGE run time, to half ( 50% )
* average of 1000 different cases, with different values of (SESSION) variables to be tested

I wouldnt say this would impossible .. if put work into optimizing your script

/halojoy
2005-07-18
sweden

[halojoy]

now

Tim, I saw you were here posting in some other topic.

I want to know:
How much faster did you get your script?
% percent, please

My advices are not bad, I think.
It is not difficult to measure page execution time.
I am sure you know howto.
🆒

[TimTimTimma]

renaming the sessions has had no effect :-(, and I am still trying to weed through the if statements

[Drakla]

Renaming the values wouldn't have any effect, especially if you were using microtime to check execution speed as it's only the parsing stage that cares about that, and that happens before any timing in the script is done.

You could also try doing as few output statements as possible, concatenating the output before echoing it, or by using buffering, but if you're actually noticing the speed difference at a perceptual level then that would be miniscule so it sounds like you've got some heavy logic going on somewhere inside.

Can't help any more as I'm a complete baby about code formatting and highlighting so I can't look at your code. Sorry dude.

[thorpe]

looking at your code.... i dont really understand why you have so much stored in sessions. do they really need to be in there? im sure its different with php, but i know with asp sessions are stored in memory, and should therefore be used sparingly. of course, this is php, and im sure things are handled better, but personlly, i never store much more than a userID in a session.

[halojoy]

thanks for comments

Drakla and thorpe
I was not quite right when suggesting a simple re-naming SESSION variables

Personally I like small files.
as loading time for an 20kB php script and checking syntax of it
would be like twice as long time
as
as loading time for an 10kB php script and checking syntax of it

at least smaller HTML pages, with image.gif is faster to serve than big, big image.bmp

while renaming varibles like
$SESSION['addddrt'][$SESSION['arthhyy']]
to
$arty
I even think I would be able to shrink this include PHP to 1/3 ( 33% )

Yes, that is what think is possible here. From 19.100 bytes ro like 6.400 bytes.

Now, I suspect like Drakla, there is some hidden 'slow down thing here,
and you know me:
I would strongly suspect mysql( functions); in this case, being the bottleneck

We have to see if someone smarter can understand what coulsd be improved with TimTimTimma script.
How about someone from your own group of phpublisher BETA-testers?

/halojoy - not smart enough with PHP to solve this one
today

.... but when tomorrow comes
.... !!!!

[TimTimTimma]

ive been inquiring everywhere, no one can figure it out, I checked with my host today, and the did a traceroute and it looks like there is an 'internet burp' somewhere along the grabing of the page from the server and then outputting to the browser. He said that the tech guys should fix this shortly, he said pages using heavy sql/ or loops will see the results of this issue, he said give it a day or two to wait for the network bandwidth tech people to fix this and try again, if the problem persists its totally script.

I personally think its the script, however ive noticed even the smallest pages are begining to take the longest time to load, and it use to only take 10 MS to load a page, now sometimes it takes 10 SECONDS! I am hoping that when this issue is resolved everything will work properly..

However i went ahead and split the file into 2 sections (index.php for the module and frontpage which made the front page load ALOT faster) and Full_Story.php, which I found was the laggy part, i disabled a few features, and the page loads 5x faster now, but is still slow as a result of what i believe to be the internet burp thing. In each initial visit the page generates the comments, and for every comment it searchs the database for the author, if it finds the author it retusn their signature, avatar, and post count. I noticed then that if there were more than 5 comments the page loaded VERY slow, so I added some pagination, and it is still slow unfortunetly :-( so i disabled the signature, post count, and avatar features, and the pages are much faster now, I am fairly certain this is obviously the culprete for slow page generation, but I would very much enjoy have made a CMS that has signatures avatars and post counts :-), is there any easier, less stressful way to accomplish this, without having to sacrifice page generation time?

here is a link to the new seperate file: http://phpublisher.net/code/Full_Story.txt

[TimTimTimma]

BUMP

Cmon, I could use alittle bit of help here :-(