Colin Watson wrote: > Oh, and have we done any benchmarking on the Ubuntu kernel with PAE > turned off vs. on? http://article.gmane.org/gmane.linux.kernel/845637 > (parent: http://article.gmane.org/gmane.linux.kernel/845514, thread: > http://thread.gmane.org/gmane.linux.kernel/843866) has some fairly > excoriating comments, noting that it slows down little things like > fork() fairly significantly. >
Anyone running the Jaunty server kernel already suffers the PAE performance hit, though I have no idea how significant that might be. > It makes sense to me to offer a clear choice between non-PAE and PAE (I > certainly understand why people might well not want to jump to amd64 > given the pain around compatibility libraries, although note that we're > actually putting serious effort into multiarch now), and I have no > issues with doing that; it'll help some users and satisfy OEMs. I just > question whether we should be setting PAE as the default. > I'm fine with offering users a choice. Given that we're not going to have runtime PAE detection in Karmic, I consider this an installer implementation issue. rtg -- Tim Gardner [email protected] -- Ubuntu-installer mailing list [email protected] Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-installer
