site_archiver@lists.apple.com Delivered-To: darwin-dev@lists.apple.com Peter Seebach writes:

In message <74ED7582-F08C-4A48-A3FB-B250951A1E62@elis.ugent.be>, Jonas Maebe wr ites:

...

If it only hurt performance a bit I would never have changed the code from using fork (which is what we used on all other *nix ports) into using vfork in the first place. But a 25% to 40% slowdown caused by 173 system calls in the process of compiling about 180 kloc is astronomical in my view.

It does seem unusually large, and I'd be interested in seeing what caused it.

There was an interesting fork/exec performance thread quite a while back. (http://lists.apple.com/archives/darwin-kernel/2002/Sep/msg00060.html) I see they've fixed the underlying problem in this particular thread (which turned out to be that the fork/exec cost scaled poorly with stack size), but shell scripts (expecially configure) remain painfully slow compared to Linux or Solaris on the same hardware. Lmbench still shows fork+exec taking roughly 5x-10x as long on MacOSX as Linux on the same hardware. Drew _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.appl... This email sent to site_archiver@lists.apple.com