NSImageView Draw Performance

[Andreas Höschler]

Hi all,

I have developed an app that draws OSM maps like so:

NSImageView subclass:
==========================
- (void)drawRect:(NSRect)rect
{
      [[NSColor whiteColor] set];
      NSRectFill(rect);   

      NSBezierPath *path = [mapPath bezierPath];
      ...
      [path fill];
      ...

      NSBezierPath *path = [mapPath bezierPath];
      ...
      [path stroke];
      ...
}

On my development machine

System Configuration: SUN MICROSYSTEMS SUN FIRE X4170 SERVER          
Intel(R) Xeon(R) CPU           E5540  @ 2.53GHz CPU 1
Intel(R) Xeon(R) CPU           E5540  @ 2.53GHz CPU 2

drawing the map/image (execution of the drawRect: code) is ultra fast (no delay 
at all).

However, on the production system

System Configuration: Sun Microsystems     Sun Fire X4240 (2.7GHz)
Quad-Core AMD Opteron(tm) Processor 2384 CPU 1
Quad-Core AMD Opteron(tm) Processor 2384 CPU 2

drawing the image (execution of the drawRect: code) takes 3-4s! :-( I do not 
want to believe that the Xenon CPU is so much faster and rather suppose a 
difference in the set of installed (used) libraries!? I use

        rsync -avz -e ssh /opt/GNUstep root@<target machine>:/opt

to deploy my software. So at least everything below /opt/GNUstep must be 
identical!?

I am clueless!! What can I do to find out which non-GNUstep libs responsible 
for drawling are used and why drawing takes so long on the target machine? Any 
ideas?

Thanks a lot in advance,

  Andreas



BTW: My GNUstep tree is a couple of years old:

        cd back
        ./configure
        make
        make install