RE: [CFCDev] Scoping Error

[Roland Collins]

Yep - which is why I said it would be a rare and application-dependent thing
;)

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf
Of Jim Davis
Sent: Thursday, November 10, 2005 7:59 PM
To: CFCDev@cfczone.org
Subject: RE: [CFCDev] Scoping Error

From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf
Of Roland Collins
Sent: Thursday, November 10, 2005 3:38 PM
To: CFCDev@cfczone.org
Subject: RE: [CFCDev] Scoping Error

>We've already been over this, but this can have performance implications
when used with highly iterative operations.  >Consider this (contrived, but
simple) example:


I agree. but have you actually tested this?  The performance difference is
REALLY small.

Using your code I built one function which uses the function local, one
which uses a directly declared function local variable and one which just
returns the equation.  I wrapped each in a getTickCount() and display the
results.  Here are some run times (I threw away the first one to allow for
compile time):

Local Time = 230
No Local Time = 200
No Local, No Middle-man Time = 201

Local Time = 270
No Local Time = 240
No Local, No Middle-man Time = 181

Local Time = 280
No Local Time = 210
No Local, No Middle-man Time = 191

Local Time = 221
No Local Time = 190
No Local, No Middle-man Time = 150

Local Time = 220
No Local Time = 180
No Local, No Middle-man Time = 170

Local Time = 210
No Local Time = 170
No Local, No Middle-man Time = 180

Local Time = 220
No Local Time = 191
No Local, No Middle-man Time = 160

So, yes, using the local was ALWAYS slower... but by only by a (very) small
fraction of ms per iteration (remember these tests for 10,000 iterations
each).  (Mine isn't a very fast research server either.)

I honestly don't think you'd ever be able to reliably measure a difference
in a real application.

Pretty much all techniques for improving development results in slower code
(although it's hard to see after 10 years of optimization the OO paradigm
itself is "slow" compared to straight procedural code for the same tasks).
It's a pretty reliable rule of thumb that anything that makes code easier to
write and easier to maintain will make code slower.  ;^)

Here's the code I used (not the best test, of course):

<cfscript>
function addA(a, b) {
        var local = StructNew();
        local.result = arguments.a + arguments.b;
        return local.result;
}

function addB(a, b) {
        var result = arguments.a + arguments.b;
        return result;
}

function addC(a, b) {
        return arguments.a + arguments.b;
}
</cfscript>
 
<cfset bTime = getTickCount() />
<cfloop from="1" to="10000" index="i">
        <cfset result = addA(i, 100)>
</cfloop>
<cfset eTime = getTickCount() />
<cfoutput>Local Time = #eTime - bTime#<br></cfoutput>
 
<cfset bTime = getTickCount() />
<cfloop from="1" to="10000" index="i">
        <cfset result = addB(i, 100)>
</cfloop>
<cfset eTime = getTickCount() />
<cfoutput>No Local Time = #eTime - bTime#<br></cfoutput>
 
<cfset bTime = getTickCount() />
<cfloop from="1" to="10000" index="i">
        <cfset result = addC(i, 100)>
</cfloop>
<cfset eTime = getTickCount() />
<cfoutput>No Local, No Middle-man Time = #eTime - bTime#<br></cfoutput>

Jim Davis




----------------------------------------------------------
You are subscribed to cfcdev. To unsubscribe, send an email to
cfcdev@cfczone.org with the words 'unsubscribe cfcdev' as the subject of the
email.

CFCDev is run by CFCZone (www.cfczone.org) and supported by CFXHosting
(www.cfxhosting.com).

An archive of the CFCDev list is available at
www.mail-archive.com/cfcdev@cfczone.org







----------------------------------------------------------
You are subscribed to cfcdev. To unsubscribe, send an email to 
cfcdev@cfczone.org with the words 'unsubscribe cfcdev' as the subject of the 
email.

CFCDev is run by CFCZone (www.cfczone.org) and supported by CFXHosting 
(www.cfxhosting.com).

An archive of the CFCDev list is available at 
www.mail-archive.com/cfcdev@cfczone.org