Win32 Performance

Ruslan Zasukhin sunshine at public.kherson.ua
Fri Mar 21 16:39:10 CST 2003


on 3/21/03 4:42 AM, Brendan Murphy at bmurf at comcast.net wrote:

Hi Brendan,

Interesting results.

I see only few explanations of this:

1) this is TRUE

2) V4RB WIN is compiled by CodeWarrior.
   Users of Valentina for Director (made by Visual C++) say they do not have
such problems....So if this point is true, then sounds like codeWarrior
generate code 8 times slower of Visual.
   I hardly believe in this, because CodeWarrior, as I recall, claim that it
generate code better of Visual.

3) ??? Right now I do not have more ideas...
    I will need future do some tests and profiling to see what is wrong and
where...    


> I am using REALbasic 4.5.3 and V4RB 1.9.8b1 and I am seeing
> performance problems on the Win32 side.
> 
> In my program I have a method that will generate database
> entries and I have optimized it so it only concentrating on
> database activities. When I run this method, I see
> performance problems with the PC version of my program.
> 
> My original test showed the following times for adding
> 10,000 records to the database.
> 
>    Mac OS X: 18.22 seconds
> 
>    PC Win XP: 1171.58 seconds
>     (yes that is over a thousand seconds!)
> 
> After examining the underlying code I found the problem. I
> was calling flush in the intermediate code which meant it
> was calling flush approximately 10,000 times.
> 
> After removing the unnecessary flush calls, I got the
> following results.
> 
>    Mac OS X: 1.6 seconds
> 
>    PC Win XP: 12.63 seconds
> 
> Now these numbers are reasonable.
> 
> For comparison sake, here are the speeds of my test
> computers.
> 
>    Mac: 1 Ghz DP
> 
>    PC: 1.2 Ghz
> 
> Yet, the PC version is still 8x slower than my OS X version.
> Now I really like my Mac, but I think I would be a fool to
> conclude that my Mac is 8x faster than my PC! So, there must
> be another explanation.
> 
> Are there any known performance problems in the Win32 side
> of things? Something isn't right here with almost an order
> of magnitude of order of difference in execution times.


-- 
Best regards,
Ruslan Zasukhin      [ I feel the need...the need for speed ]
-------------------------------------------------------------
e-mail: ruslan at paradigmasoft.com
web: http://www.paradigmasoft.com

To subscribe to the Valentina mail list go to:
http://listserv.macserve.net/mailman/listinfo/valentina
-------------------------------------------------------------



More information about the Valentina mailing list