Re: trying to understand poor performance of make + cygwin on W2K

[Rob Tulloh]

Please note that I found that Linux blows the doors off
of our builds on Solaris, AIX, HP/UX, and Win2K.
I see 2x faster times when building on Red Hat 6.2 on
i686 platforms. Maybe your problem is related to the platform and not the
tool?

You might also like to try the native Win32 port of GNU
make to see if that helps. Emulation through cygwin has
some cost associated with it. If you need the performance
boost, put the native port in and use your Cygwin tools with
it. Worth a try if you can afford the time to get it set up.Of course,
if you need all the new features of GNU make (e.g. the parallism
feature that Eli mentions), you are better off to stay with the Cygwin
version of the tool.

My $.02,

Rob

Eli Zaretskii wrote:

> On Tue, 30 Apr 2002, Ken Faiczak wrote:
>
> > If I completely build our tree (about
> > then rerun the make from the top it takes 9.5 seconds on linux
> >
> > if I do the same test on the same machine running
> > (machine P3-500 512MB)
> > win2k +cygwin 1.3.9 +make (3.79.1)
> > it takes 3.5 minutes
> >
> > so 210 seconds versus 9 seconds.
> > all its doing is recursing down the tree, testing the dependancies
> > and determining it has nothing to do, so its not compiling anything
> > its all make +cygwin, I think (ie no gcc invoked anywhere)
> >
> > any ideas on what to try?
> > is this an issue with the cygwin fork() implementation??
>
> Does the Cygwin port supports parallelism?  I suggest to try a
> non-parallel run (by adding .NOTPARALLEL target to the Makefile's), and
> see if the results change.
>
> It could also be an issue with fork.  The Cygwin mailing list is probably
> a better place to ask about this.
>
> _______________________________________________
> Make-w32 mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/make-w32

tulloh.vcf
Description: Card for Rob Tulloh