Linux build times, and VTK optimized builds

[Brent Burton]

On Mon, 29 Nov 1999, Bryan VanDeVen wrote:
> > My installation is on PII 400 with SCSI-2 and a S3 video card. It took over
> > 12 hours to compile. Since some reported good result with AMD K3, I  went
> > out and get a AMD chip to see if it cut down the compilation time. I must
> > not done the switches right, it still took over 11 hours.
> 
> Well, VTK takes a good three or four hours to build on my dual PII450 with
> 256MB.  So 11 to 12 hours on the above setup does not seem unreasonable.
> If you have plenty of ram you might consider using the "-j" option to
> initiate a parallel make.  Obviously with one processor you won't truly
> get a parallel build but some of the disk IO can then be done 
> asynchronously so you might see a small performance bump from that.

Whoa, 3 to 4 hours seems unreasonable on such a machine.

I didn't remember VTK taking so long to build and just last week I had to
build VTK twice in one day.  Just now I cleaned out all object files and
*.so libraries and rebuilt.  The entire VTK 2.4 built in 30 minutes.  
libVTKCommon.so was built in 2 minutes, and libVTKCommonTcl.so was built
in two more minutes.  Make compiles a new file almost every second or two.

This is on a 400MHz P-III (they don't make 400, this is an old
prototype chip from Intel I got before their release) with 128MB RAM
and IDE disk running Redhat 5.1 (kernel 2.0.34).  I run a serial make
to build common, graphics, imaging and patented.  Obviously, it's
neither memory- nor disk-bound.

So, what are your machines doing that takes a build so long?  There's
not that much of a difference between the P-II and P-III raw performance
to warrant such a large compile time difference.


On a related note, when I specify optimization flags (-O or -O2)
in the user.make file, VTK crashes in vtkMarchingCubes::Execute().
When optimizations are not enabled, it does not crash.  I haven't
been able to pin down the exact cause, but I think it is with
the templates used here.  This is on Redhat 5.1, GCC 2.7.2.3.
Has anybody else run into this?

-Brent





-----------------------------------------------------------------------------
This is the private VTK discussion list.  Please keep messages on-topic.
Check the FAQ at: <http://www.automatrix.com/cgi-bin/vtkfaq>
To UNSUBSCRIBE, send message body containing "unsubscribe vtkusers" to
<majordomo at gsao.med.ge.com>.  For help, send message body containing
"info vtkusers" to the same address.     Live long and prosper.
-----------------------------------------------------------------------------