optimizing code

[Andreas Paasch]

I've finally gotten my nice little system working and it's gone live.
Now, I spent time optimizing my code and adding a little smart functionality
here and there, based on needs and simplicity.

In this context, I was wondering about something.

I have a growing include.inc file that holds all my functions in it.

Thinking speed, I was thinking that it might be a bit faster to split that
include.inc file up into the different functions and then only include_once
those that actually are needed by the different pages.
But without any reference and in-depth knowledge, I don't know if this is a
win situation or loose, in matters of time.

I'm running Mandrake 9.2 here on a 2.4 kernel but was at the same time
considering to upgrade to 2.6 kernel, once it seems to be solid enough.
According to the NG Mandrake 2.6 kernel still has a very few issues that I'd
like to avoid.

But considering optimization and time, would it be most practical and faster
to split my functions or keep it as is?

/Andreas
--
Registeret Linux user #292411

[Jochen Daum]

Re: optimizing code

Hi Andreas!

On Sun, 15 Feb 2004 13:24:30 +0100, Andreas Paasch
<Andreas@Paasch .Net> wrote:
[color=blue]
>
>I've finally gotten my nice little system working and it's gone live.
>Now, I spent time optimizing my code and adding a little smart functionality
>here and there, based on needs and simplicity.
>
>In this context, I was wondering about something.
>
>I have a growing include.inc file that holds all my functions in it.
>
>Thinking speed, I was thinking that it might be a bit faster to split that
>include.inc file up into the different functions and then only include_once
>those that actually are needed by the different pages.
>But without any reference and in-depth knowledge, I don't know if this is a
>win situation or loose, in matters of time.
>
>I'm running Mandrake 9.2 here on a 2.4 kernel but was at the same time
>considering to upgrade to 2.6 kernel, once it seems to be solid enough.
>According to the NG Mandrake 2.6 kernel still has a very few issues that I'd
>like to avoid.
>
>But considering optimization and time, would it be most practical and faster
>to split my functions or keep it as is?[/color]

I have measured this for several thousand lines of array declarations,
and it is faster than most short loops, that I run. Its not worth
optimising, only optimise the slowest code.

HTH, Jochen
--
Jochen Daum - Cabletalk Group Ltd.
PHP DB Edit Toolkit -- PHP scripts for building
database editing interfaces.

http://sourceforge.net/projects/phpdbedittk/

[CountScubula]

Re: optimizing code


Just a suggestion, it the difference is slight, and you have few vistitors
(less than 1000 day) you might just want to leave ti as is for simplicity.

--
Mike Bradley
http://www.gzentools.com -- free online php tools
"Andreas Paasch" <Andreas@Paasch .Net> wrote in message
news:dvJXb.9105 0$jf4.5681483@n ews000.worldonl ine.dk...[color=blue]
>
> I've finally gotten my nice little system working and it's gone live.
> Now, I spent time optimizing my code and adding a little smart[/color]
functionality[color=blue]
> here and there, based on needs and simplicity.
>
> In this context, I was wondering about something.
>
> I have a growing include.inc file that holds all my functions in it.
>
> Thinking speed, I was thinking that it might be a bit faster to split that
> include.inc file up into the different functions and then only[/color]
include_once[color=blue]
> those that actually are needed by the different pages.
> But without any reference and in-depth knowledge, I don't know if this is[/color]
a[color=blue]
> win situation or loose, in matters of time.
>
> I'm running Mandrake 9.2 here on a 2.4 kernel but was at the same time
> considering to upgrade to 2.6 kernel, once it seems to be solid enough.
> According to the NG Mandrake 2.6 kernel still has a very few issues that[/color]
I'd[color=blue]
> like to avoid.
>
> But considering optimization and time, would it be most practical and[/color]
faster[color=blue]
> to split my functions or keep it as is?
>
> /Andreas
> --
> Registeret Linux user #292411[/color]

[Andreas Paasch]

Re: optimizing code

CountScubula wrote:
[color=blue]
>
> Just a suggestion, it the difference is slight, and you have few vistitors
> (less than 1000 day) you might just want to leave ti as is for simplicity.
>
> --
> Mike Bradley
> http://www.gzentools.com -- free online php tools[/color]

[snip]

Thanks, I'll leave it as it is then ...

/Andreas
--
Registeret Linux user #292411

[Andreas Paasch]

Re: optimizing code

Jochen Daum wrote:
[color=blue]
> Hi Andreas!
>[/color]

[snip]
[color=blue]
> I have measured this for several thousand lines of array declarations,
> and it is faster than most short loops, that I run. Its not worth
> optimising, only optimise the slowest code.
>
> HTH, Jochen[/color]

Thanks, I'll leave it as it is :-)

--
Registeret Linux user #292411

[David Mackenzie]

Re: optimizing code

On Mon, 16 Feb 2004 08:39:31 +1300, Jochen Daum
<jochen.daum@ca ns.co.nz> wrote:
[color=blue]
>Hi Andreas!
>
>On Sun, 15 Feb 2004 13:24:30 +0100, Andreas Paasch
><Andreas@Paasc h.Net> wrote:
>[color=green]
>>
>>I've finally gotten my nice little system working and it's gone live.
>>Now, I spent time optimizing my code and adding a little smart functionality
>>here and there, based on needs and simplicity.
>>
>>In this context, I was wondering about something.
>>
>>I have a growing include.inc file that holds all my functions in it.
>>
>>Thinking speed, I was thinking that it might be a bit faster to split that
>>include.inc file up into the different functions and then only include_once
>>those that actually are needed by the different pages.
>>But without any reference and in-depth knowledge, I don't know if this is a
>>win situation or loose, in matters of time.
>>
>>I'm running Mandrake 9.2 here on a 2.4 kernel but was at the same time
>>considering to upgrade to 2.6 kernel, once it seems to be solid enough.
>>According to the NG Mandrake 2.6 kernel still has a very few issues that I'd
>>like to avoid.
>>
>>But considering optimization and time, would it be most practical and faster
>>to split my functions or keep it as is?[/color]
>
>I have measured this for several thousand lines of array declarations,
>and it is faster than most short loops, that I run. Its not worth
>optimising, only optimise the slowest code.[/color]

In my experience, the time taken to open and close several include
files is much greater than parsing one large file.

--
David ( @priz.co.uk )

[Andreas Paasch]

Re: optimizing code

David Mackenzie wrote:
[color=blue]
> On Mon, 16 Feb 2004 08:39:31 +1300, Jochen Daum
> <jochen.daum@ca ns.co.nz> wrote:
>[color=green]
>>Hi Andreas!
>>
>>On Sun, 15 Feb 2004 13:24:30 +0100, Andreas Paasch
>><Andreas@Paas ch.Net> wrote:
>>[color=darkred]
>>>
>>>I've finally gotten my nice little system working and it's gone live.
>>>Now, I spent time optimizing my code and adding a little smart
>>>functionalit y here and there, based on needs and simplicity.
>>>
>>>In this context, I was wondering about something.
>>>
>>>I have a growing include.inc file that holds all my functions in it.
>>>
>>>Thinking speed, I was thinking that it might be a bit faster to split
>>>that include.inc file up into the different functions and then only
>>>include_on ce those that actually are needed by the different pages.
>>>But without any reference and in-depth knowledge, I don't know if this is
>>>a win situation or loose, in matters of time.
>>>
>>>I'm running Mandrake 9.2 here on a 2.4 kernel but was at the same time
>>>considerin g to upgrade to 2.6 kernel, once it seems to be solid enough.
>>>According to the NG Mandrake 2.6 kernel still has a very few issues that
>>>I'd like to avoid.
>>>
>>>But considering optimization and time, would it be most practical and
>>>faster to split my functions or keep it as is?[/color]
>>
>>I have measured this for several thousand lines of array declarations,
>>and it is faster than most short loops, that I run. Its not worth
>>optimising, only optimise the slowest code.[/color]
>
> In my experience, the time taken to open and close several include
> files is much greater than parsing one large file.
>[/color]

That is what came to my mind too, after reading the other two posts.

/Andreas
--
Registeret Linux user #292411

[Perttu Pulkkinen]

Re: optimizing code


"Andreas Paasch" <Andreas@Paasch .Net> kirjoitti viestissä
news:dvJXb.9105 0$jf4.5681483@n ews000.worldonl ine.dk...
....[color=blue]
> I have a growing include.inc file that holds all my functions in it.[/color]
....[color=blue]
> But considering optimization and time, would it be most practical and[/color]
faster[color=blue]
> to split my functions or keep it as is?[/color]

I have not measured any times, I think that is not the problem usually.
But in my experience, this division is VERY useful:

- database library (general database things and/or wrapping of certain
database system (like mysqsl) functions into more general db_functions)
- standard library (your own often used string and html functions f.ex.)
- image library (handling of images adn thunmbnails)
- userlibrary (userside functions of certain application)
- adminlibrary (admin sideof certain application)

Advantages:

* Debugging and developing the code is easier when divided in this way.
* If first three are wellmade you can use them in almost every application
without big changes.
* When user and admin functions are in separate librarys, problems in
another don't cause the crash of another.

-perttu, jkl, finland