Timer

[scuniverse@gmail.com]

I've made a script that is called every hour from a cron web service,
the problem is that it takes like 50s to execute it and when I run it
manually (just by putting the url in the browser) it takes less than 1s
to execute it.

In this script there's a while loop and for testing proposes I included
a timer in it to see how much time each loop takes.

With the web cron service we have
[1] => 0.042862
[2] => 0.046418
[3] => 0.049904
[4] => 21.338746
[5] => 21.340762
[6] => 52.182604
[7] => 52.738617
[8] => 52.887804

Manually we have
[1] => 0.133298
[2] => 0.144081
[3] => 0.152357
[4] => 0.155286
[5] => 0.15921
[6] => 0.162126
[7] => 0.184527
[8] => 0.20398

The jump from the loop 3 to 4 and from 5 to 6 in the web cron is
constant and happens all the time, when I do it manually I don't see
anything like that.

Any ideas why it runs differently if using the web service and doing it
manually ?

[Jerry Stuckle]

Re: Timer

scuniverse@gmai l.com wrote:[color=blue]
> I've made a script that is called every hour from a cron web service,
> the problem is that it takes like 50s to execute it and when I run it
> manually (just by putting the url in the browser) it takes less than 1s
> to execute it.
>
> In this script there's a while loop and for testing proposes I included
> a timer in it to see how much time each loop takes.
>
> With the web cron service we have
> [1] => 0.042862
> [2] => 0.046418
> [3] => 0.049904
> [4] => 21.338746
> [5] => 21.340762
> [6] => 52.182604
> [7] => 52.738617
> [8] => 52.887804
>
> Manually we have
> [1] => 0.133298
> [2] => 0.144081
> [3] => 0.152357
> [4] => 0.155286
> [5] => 0.15921
> [6] => 0.162126
> [7] => 0.184527
> [8] => 0.20398
>
> The jump from the loop 3 to 4 and from 5 to 6 in the web cron is
> constant and happens all the time, when I do it manually I don't see
> anything like that.
>
> Any ideas why it runs differently if using the web service and doing it
> manually ?
>[/color]

Not without knowing what the script is doing!

--
=============== ===
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex@attgl obal.net
=============== ===

[lorento]

Re: Timer

How you cron it?

Are you using wget? or execute the php script from unix shell?

I think its better use the wget command instead of running from unix
shell.

Regards,

Lorento
--

http://www.mastervb.net

http://www.immersivelounge.com

[Jerry Stuckle]

Re: Timer

lorento wrote:[color=blue]
> How you cron it?
>
> Are you using wget? or execute the php script from unix shell?
>
> I think its better use the wget command instead of running from unix
> shell.
>
> Regards,
>
> Lorento
> --
> http://www.mastervb.net
> http://www.immersivelounge.com
>[/color]

The same way you set up any other cron - in the Unix shell.

And yes, you can easily execute a PHP script from the shell - or, in Windows,
the command line.

--
=============== ===
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex@attgl obal.net
=============== ===

[scuniverse@gmail.com]

Re: Timer

The script is rather big... there's like 600 lines, however it's quite
simple.

At the beggining there's a select mysql query and afterwars a while
loop, on the loop there's 3 possible cases and each one has 2 select
and 3 update mysql query, in between the 2 select and 3 updates there's
a bunch of basic maths and that's pretty much everything.

When I run it manually it output a lot of division by zero errors but
no other error

[Metin Savignano]

Re: Timer

[color=blue]
> When I run it manually it output a lot of division by zero errors but
> no other error[/color]

Please specify what you mean by "manually": Do you run it in the shell?
Or do you run it through Apache (i.e. by a page request)?

If run from the cron: Is it initiated in exactly the same way?

If you start it manually, do you do this locally or remote?

Guess: If you get a lot of division by zero errors (which is very
unclean, BTW), it might be that the script is run with different PHP
settings. This again may be due to the fact that you start it in
different ways.

You might put a phpinfo() in your script and compare the output of both
ways you start it.

Just some guesses. We need more info to be more specific.

Metin

[scuniverse@gmail.com]

Re: Timer

When I said manually I meant by a page request, the script is in a
webserver and I request by putting the URL in the browser and the
webcron request the same page.

You said it wasnt clean the warnings about the division by zero so I
set the error reporting to a lower value (E_ERROR) and it fixed
everything, now I get the same times on both cases :D I dont really
understand the why though.

[Metin Savignano]

Re: Timer

> When I said manually I meant by a page request, the script is in a[color=blue]
> webserver and I request by putting the URL in the browser and the
> webcron request the same page.[/color]

Does the cron run on the same machine as the web server?
[color=blue]
> You said it wasnt clean the warnings about the division by zero so I
> set the error reporting to a lower value (E_ERROR) and it fixed
> everything, now I get the same times on both cases :D I dont really
> understand the why though.[/color]

Even though this is not what you wanted, it means we get closer to the
problem. Are you sure the script had run successfully till its end
before you fixed these problems? Maybe it was canceled early, so it was
faster because of that...

Next step would be: Write to a log file every few lines of your script,
so you can find out where in the script this delay happens, and then
take a closer a look at it.

Or post it, and let us take a closer look together.

[Jerry Stuckle]

Re: Timer

scuniverse@gmai l.com wrote:[color=blue]
> The script is rather big... there's like 600 lines, however it's quite
> simple.
>
> At the beggining there's a select mysql query and afterwars a while
> loop, on the loop there's 3 possible cases and each one has 2 select
> and 3 update mysql query, in between the 2 select and 3 updates there's
> a bunch of basic maths and that's pretty much everything.
>
> When I run it manually it output a lot of division by zero errors but
> no other error
>[/color]

Well, a lot of division by zero errors tells me there's a lot of sloppy
programming there, and anything could be happening.

But in addition to that - there is no way anyone can tell you what's going on
without detailed knowledge of what you're doing in the loops. I suggest you cut
down the amount of code in the slow loops until you locate the specific line(s)
of code causing the problem.

--
=============== ===
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex@attgl obal.net
=============== ===

[Jerry Stuckle]

Re: Timer

scuniverse@gmai l.com wrote:[color=blue]
> When I said manually I meant by a page request, the script is in a
> webserver and I request by putting the URL in the browser and the
> webcron request the same page.
>
> You said it wasnt clean the warnings about the division by zero so I
> set the error reporting to a lower value (E_ERROR) and it fixed
> everything, now I get the same times on both cases :D I dont really
> understand the why though.
>[/color]

Setting the warnings to a lower value isn't fixing anything. It's just hiding
the errors you have. You still have crappy code in there.

--
=============== ===
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex@attgl obal.net
=============== ===

[scuniverse@gmail.com]

Re: Timer

Here is part of the code, it's a little long though

<?php
//Timer
function getmicrotime()
{
list($usec, $sec) = explode(" ", microtime());
return((float)$ usec + (float)$sec);
}
$time_start = getmicrotime();

function totaltime($time _start)
{
$time_end = getmicrotime();
return(round(($ time_end - $time_start), 6));
}

//Code
include_once(". ./inc/config.php");
error_reporting (E_ERROR);

dbConnect("sess ions");
$sqlUID = "SELECT ID, Race, Min, Gaz, Size FROM user";
$UsersID = mysql_query($sq lUID) or die("Query failed");

$i = 1;
while ($UserID = mysql_fetch_arr ay($UsersID)) {
$UrsID = $UserID['ID'];

// Terran
if ($UserID['Race'] == 'Terran') {
$sqlOBu = "SELECT * FROM units WHERE ID = '$UrsID'";
$UserOnB = mysql_query($sq lOBu) or die("Query B failed");
$UserONBU = mysql_fetch_arr ay($UserOnB);

$sqlBU = "SELECT * FROM build WHERE ID = '$UrsID'";
$UserB = mysql_query($sq lBU) or die("Query B failed");
$UserBU = mysql_fetch_arr ay($UserB);

//units build on the barracks
$FACB = $UserONBU[4] + $UserONBU[6] + $UserONBU[8] +
$UserONBU[10];
//units build on the factory
$FACF = $UserONBU[12] + $UserONBU[14] + $UserONBU[16];
//Buildings built by the SCV
$FACSCV = $UserBU[2] + $UserBU[4] + $UserBU[6] + $UserBU[8] +
$UserBU[10] + $UserBU[12] + $UserBU[14] + $UserBU[16] + $UserBU[18] +
$UserBU[20];

//SCV per tick
$SCVPT = $UserBU[1] * 5;
//Total SCV built this tick
$SCVTA = $UserONBU[2] - $SCVPT;
//same as for SCV only for the rest of the units
$MarPT = $UserBU[9] * 4.2 * ($UserONBU[6] / $FACB);
$MarTA = $UserONBU[6] - $MarPT;
$FirPT = $UserBU[9] * 4.2 * ($UserONBU[8] / $FACB);
$FirTA = $UserONBU[8] - $FirPT;
$GhoPT = $UserBU[9] * 2 * ($UserONBU[10] / $FACB);
$GhoTA = $UserONBU[10] - $GhoPT;
$VulPT = $UserBU[17] * 3.3 * ($UserONBU[12] / $FACF);
$VulTA = $UserONBU[12] - $VulPT;
$GolPT = $UserBU[17] * 2.5 * ($UserONBU[14] / $FACF);
$GolTA = $UserONBU[14] - $GolPT;
$TanPT = $UserBU[17] * 2 * ($UserONBU[16] / $FACF);
$TanTA = $UserONBU[16] - $TanPT;


//Total SCVs and remaining SCVs to be built
if ($SCVTA < 0 or $SCVTA == 0) {
$SCVTAD = $UserONBU[2] + $UserONBU[1];
$SCVOB = 0;
} else {
$SCVTAD = $SCVPT + $UserONBU[1];
$SCVOB = $SCVTA;
}
if ($MarTA < 0 or $MarTA == 0) {
$MarTAD = $UserONBU[6] + $UserONBU[5];
$MarOB = 0;
} else {
$MarTAD = $MarPT + $UserONBU[5];
$MarOB = $MarTA;
}
if ($FirTA < 0 or $FirTA == 0) {
$FirTAD = $UserONBU[8] + $UserONBU[7];
$FirOB = 0;
} else {
$FirTAD = $FirPT + $UserONBU[7];
$FirOB = $FirTA;
}
if ($GhoTA < 0 or $GhoTA == 0) {
$GhoTAD = $UserONBU[10] + $UserONBU[9];
$GhoOB = 0;
} else {
$GhoTAD = $GhoPT + $UserONBU[9];
$GhoOB = $GhoTA;
}
if ($MedTA < 0 or $MedTA == 0) {
$MedTAD = $UserONBU[3] + $UserONBU[4];
$MedOB = 0;
} else {
$MedTAD = $MedPT + $UserONBU[3];
$MedOB = $MedTA;
}
if ($VulTA < 0 or $VulTA == 0) {
$VulTAD = $UserONBU[11] + $UserONBU[12];
$VulOB = 0;
} else {
$VulTAD = $VulPT + $UserONBU[11];
$VulOB = $VulTA;
}
if ($GolTA < 0 or $GolTA == 0) {
$GolTAD = $UserONBU[13] + $UserONBU[14];
$GolOB = 0;
} else {
$GolTAD = $GolPT + $UserONBU[13];
$GolOB = $GolTA;
}
if ($TanTA < 0 or $TanTA == 0) {
$TanTAD = $UserONBU[15] + $UserONBU[16];
$TanOB = 0;
} else {
$TanTAD = $TanPT + $UserONBU[15];
$TanOB = $TanTA;
}


//Total strenght
$Nstrenght = ($UserID['Size'] / 2) + ($UserBU[1] * 2 +
$UserBU[3] + $UserBU[5] + $UserBU[7] + $UserBU[9] + $UserBU[11] +
$UserBU[13] + $UserBU[15]) + $UserBU[17] + $UserBU[19] + ($UserONBU[1]
* 2 + $UserONBU[3] + $UserONBU[5] + $UserONBU[7] + $UserONBU[9] +
$UserONBU[11] + $UserONBU[13] + $UserONBU[15]);

//calcul of the SCV and command center effect on minerals and
gaz gathering
$LCC = log10(($UserBU[1] / $UserID['Size']) * 10 + 1);
$LSC = log(($UserONBU[1] / $UserID['Size']) + 1, 2);
$Fac = $LCC * $LSC * $UserID['Size'] * 100;

$NMin = (int)($Fac + $UserID[1]);
$NGz = (int)(($Fac / 3) + $UserID[2]);

//Automatic increase of land... for now.
$NSize = ($UserID['Size'] * 0.01 + 1) + $UserID['Size'];

$UpdateOB = mysql_query("UP DATE units SET
SCVEd='$SCVTAD' ,SCVIng='$SCVOB ',MarEd='$MarTA D',MarIng='$Mar OB',FireEd='$Fi rTAD',FireIng=' $FirOB'
,GhostEd='$GhoT AD',GhostIng='$ GhoOB',MedEd='$ MedTAD',MedIng= '$MedOB',VulEd= '$VulTAD',VulIn g='$VulOB',GolE d='$GolTAD',Gol Ing='$GolOB'
,TankEd='$TanTA D',TankIng='$Ta nOB' WHERE ID='$UrsID'") or
die(mysql_error ());

$UpdateBU = mysql_query("UP DATE build SET
CCB='$CCoTAD',C COB='$CCoOB',SD B='$DepTAD',SDO B='$DepOB',RefB ='$RefTAD',RefO B='$RefOB',Barr B='$BarTAD',Bar rOB='$BarOB'
,FacB='$FacTAD' ,FacOB='$FacOB' ,EBB='$BayTAD', EBOB='$BayOB',A caB='$AcaTAD',A caOB='$AcaOB',A rmB='$ArmTAD',A rmOB='$ArmOB',S FB='$SciTAD',SF OB='$SciOB'
,BunkB='$BunTAD ',BunkOB='$BunO B' WHERE ID='$UrsID'") or
die(mysql_error ());

$UpdateMG = mysql_query("UP DATE user SET
Min='$NMin',Gaz ='$NGz',NW='$Ns trenght',Size=' $NSize' WHERE
ID='$UrsID'") or die(showError(' 3', mysql_error())) ;
}



I've cut a part on the code but its basically the same thing only for
the building where here its just for the units and there's 2 more
cases afterwards for 2 other races, the division by zero errors I get
them after "//SCV per tick" and every time I divide something by
$FACB, this variable is usually not 0, its 0 now cause I'm still
testing this, that said I could put a condition there cause there's
no point to do that part when $FACB = 0 or $FACF =0.

Is it possible that the server slows down the speed of the script for
some reason ? seems to me that there's a link between the error
messages and the time it takes to run the script, and this only happens
with the web cron service and not when I request the url with a browser.

[Jerry Stuckle]

Re: Timer

scuniverse@gmai l.com wrote:[color=blue]
> Here is part of the code, it's a little long though
>
> <?php
> //Timer
> function getmicrotime()
> {
> list($usec, $sec) = explode(" ", microtime());
> return((float)$ usec + (float)$sec);
> }
> $time_start = getmicrotime();
>
> function totaltime($time _start)
> {
> $time_end = getmicrotime();
> return(round(($ time_end - $time_start), 6));
> }
>
> //Code
> include_once(". ./inc/config.php");
> error_reporting (E_ERROR);
>
> dbConnect("sess ions");
> $sqlUID = "SELECT ID, Race, Min, Gaz, Size FROM user";
> $UsersID = mysql_query($sq lUID) or die("Query failed");
>
> $i = 1;
> while ($UserID = mysql_fetch_arr ay($UsersID)) {
> $UrsID = $UserID['ID'];
>
> // Terran
> if ($UserID['Race'] == 'Terran') {
> $sqlOBu = "SELECT * FROM units WHERE ID = '$UrsID'";
> $UserOnB = mysql_query($sq lOBu) or die("Query B failed");
> $UserONBU = mysql_fetch_arr ay($UserOnB);
>
> $sqlBU = "SELECT * FROM build WHERE ID = '$UrsID'";
> $UserB = mysql_query($sq lBU) or die("Query B failed");
> $UserBU = mysql_fetch_arr ay($UserB);
>
> //units build on the barracks
> $FACB = $UserONBU[4] + $UserONBU[6] + $UserONBU[8] +
> $UserONBU[10];
> //units build on the factory
> $FACF = $UserONBU[12] + $UserONBU[14] + $UserONBU[16];
> //Buildings built by the SCV
> $FACSCV = $UserBU[2] + $UserBU[4] + $UserBU[6] + $UserBU[8] +
> $UserBU[10] + $UserBU[12] + $UserBU[14] + $UserBU[16] + $UserBU[18] +
> $UserBU[20];
>
> //SCV per tick
> $SCVPT = $UserBU[1] * 5;
> //Total SCV built this tick
> $SCVTA = $UserONBU[2] - $SCVPT;
> //same as for SCV only for the rest of the units
> $MarPT = $UserBU[9] * 4.2 * ($UserONBU[6] / $FACB);
> $MarTA = $UserONBU[6] - $MarPT;
> $FirPT = $UserBU[9] * 4.2 * ($UserONBU[8] / $FACB);
> $FirTA = $UserONBU[8] - $FirPT;
> $GhoPT = $UserBU[9] * 2 * ($UserONBU[10] / $FACB);
> $GhoTA = $UserONBU[10] - $GhoPT;
> $VulPT = $UserBU[17] * 3.3 * ($UserONBU[12] / $FACF);
> $VulTA = $UserONBU[12] - $VulPT;
> $GolPT = $UserBU[17] * 2.5 * ($UserONBU[14] / $FACF);
> $GolTA = $UserONBU[14] - $GolPT;
> $TanPT = $UserBU[17] * 2 * ($UserONBU[16] / $FACF);
> $TanTA = $UserONBU[16] - $TanPT;
>
>
> //Total SCVs and remaining SCVs to be built
> if ($SCVTA < 0 or $SCVTA == 0) {
> $SCVTAD = $UserONBU[2] + $UserONBU[1];
> $SCVOB = 0;
> } else {
> $SCVTAD = $SCVPT + $UserONBU[1];
> $SCVOB = $SCVTA;
> }
> if ($MarTA < 0 or $MarTA == 0) {
> $MarTAD = $UserONBU[6] + $UserONBU[5];
> $MarOB = 0;
> } else {
> $MarTAD = $MarPT + $UserONBU[5];
> $MarOB = $MarTA;
> }
> if ($FirTA < 0 or $FirTA == 0) {
> $FirTAD = $UserONBU[8] + $UserONBU[7];
> $FirOB = 0;
> } else {
> $FirTAD = $FirPT + $UserONBU[7];
> $FirOB = $FirTA;
> }
> if ($GhoTA < 0 or $GhoTA == 0) {
> $GhoTAD = $UserONBU[10] + $UserONBU[9];
> $GhoOB = 0;
> } else {
> $GhoTAD = $GhoPT + $UserONBU[9];
> $GhoOB = $GhoTA;
> }
> if ($MedTA < 0 or $MedTA == 0) {
> $MedTAD = $UserONBU[3] + $UserONBU[4];
> $MedOB = 0;
> } else {
> $MedTAD = $MedPT + $UserONBU[3];
> $MedOB = $MedTA;
> }
> if ($VulTA < 0 or $VulTA == 0) {
> $VulTAD = $UserONBU[11] + $UserONBU[12];
> $VulOB = 0;
> } else {
> $VulTAD = $VulPT + $UserONBU[11];
> $VulOB = $VulTA;
> }
> if ($GolTA < 0 or $GolTA == 0) {
> $GolTAD = $UserONBU[13] + $UserONBU[14];
> $GolOB = 0;
> } else {
> $GolTAD = $GolPT + $UserONBU[13];
> $GolOB = $GolTA;
> }
> if ($TanTA < 0 or $TanTA == 0) {
> $TanTAD = $UserONBU[15] + $UserONBU[16];
> $TanOB = 0;
> } else {
> $TanTAD = $TanPT + $UserONBU[15];
> $TanOB = $TanTA;
> }
>
>
> //Total strenght
> $Nstrenght = ($UserID['Size'] / 2) + ($UserBU[1] * 2 +
> $UserBU[3] + $UserBU[5] + $UserBU[7] + $UserBU[9] + $UserBU[11] +
> $UserBU[13] + $UserBU[15]) + $UserBU[17] + $UserBU[19] + ($UserONBU[1]
> * 2 + $UserONBU[3] + $UserONBU[5] + $UserONBU[7] + $UserONBU[9] +
> $UserONBU[11] + $UserONBU[13] + $UserONBU[15]);
>
> //calcul of the SCV and command center effect on minerals and
> gaz gathering
> $LCC = log10(($UserBU[1] / $UserID['Size']) * 10 + 1);
> $LSC = log(($UserONBU[1] / $UserID['Size']) + 1, 2);
> $Fac = $LCC * $LSC * $UserID['Size'] * 100;
>
> $NMin = (int)($Fac + $UserID[1]);
> $NGz = (int)(($Fac / 3) + $UserID[2]);
>
> //Automatic increase of land... for now.
> $NSize = ($UserID['Size'] * 0.01 + 1) + $UserID['Size'];
>
> $UpdateOB = mysql_query("UP DATE units SET
> SCVEd='$SCVTAD' ,SCVIng='$SCVOB ',MarEd='$MarTA D',MarIng='$Mar OB',FireEd='$Fi rTAD',FireIng=' $FirOB'
> ,GhostEd='$GhoT AD',GhostIng='$ GhoOB',MedEd='$ MedTAD',MedIng= '$MedOB',VulEd= '$VulTAD',VulIn g='$VulOB',GolE d='$GolTAD',Gol Ing='$GolOB'
> ,TankEd='$TanTA D',TankIng='$Ta nOB' WHERE ID='$UrsID'") or
> die(mysql_error ());
>
> $UpdateBU = mysql_query("UP DATE build SET
> CCB='$CCoTAD',C COB='$CCoOB',SD B='$DepTAD',SDO B='$DepOB',RefB ='$RefTAD',RefO B='$RefOB',Barr B='$BarTAD',Bar rOB='$BarOB'
> ,FacB='$FacTAD' ,FacOB='$FacOB' ,EBB='$BayTAD', EBOB='$BayOB',A caB='$AcaTAD',A caOB='$AcaOB',A rmB='$ArmTAD',A rmOB='$ArmOB',S FB='$SciTAD',SF OB='$SciOB'
> ,BunkB='$BunTAD ',BunkOB='$BunO B' WHERE ID='$UrsID'") or
> die(mysql_error ());
>
> $UpdateMG = mysql_query("UP DATE user SET
> Min='$NMin',Gaz ='$NGz',NW='$Ns trenght',Size=' $NSize' WHERE
> ID='$UrsID'") or die(showError(' 3', mysql_error())) ;
> }
>
>
>
> I've cut a part on the code but its basically the same thing only for
> the building where here its just for the units and there's 2 more
> cases afterwards for 2 other races, the division by zero errors I get
> them after "//SCV per tick" and every time I divide something by
> $FACB, this variable is usually not 0, its 0 now cause I'm still
> testing this, that said I could put a condition there cause there's
> no point to do that part when $FACB = 0 or $FACF =0.
>
> Is it possible that the server slows down the speed of the script for
> some reason ? seems to me that there's a link between the error
> messages and the time it takes to run the script, and this only happens
> with the web cron service and not when I request the url with a browser.
>[/color]

Well, among other things, when you get an error, it takes time to process the
error, log it and attempt to recover from it.

Division by zero is a hardware detected problem. So the error processing starts
at the very lowest level and has to percolate up the chain the the PHP error
recovery routines (which eventually handle it).

The bottom line is - you should never get run-time errors.

--
=============== ===
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex@attgl obal.net
=============== ===