On Mon, Sep 18, 2006 at 12:01:05PM +0200, Pavel Machek wrote: > Hi! > > > > I think performance shouldn't be hurt, since we're still batching > > > pages 1% at a time, down from 20%. A normal image has 10's of > > > thousands of pages (mine 512M laptop is ~55000 pages, for example) so > > > 1% is still~500 pages or ~2M at a time.
> Well, patch does not look too bad to me, but... > > "It is an option, so it is not important to get it right" is ugly. It > would be nice to actually benchmark this, and hardcode the option that > provides acceptable performance... I actually have a pretty bad machine wrt. writeout behavior here (hp nx5000), so i will test this. early writeout is right now about ~4-5 seconds faster (it takes ~6-8 seconds to do the final sync without early writeout). But this machine also is not really faster with compression enabled. With compression enabled, i get a flickering disk led during write, without it is solid on. Since it is a Pentium M 1400, CPU speed should be fast enough to benefit from compression. It also was always dog slow with in-kernel suspend. Short: this machine looks like it is the right candidate for this kind of benchmark, since i can probably see every small amount of improvement of writeout speed easily :-) -- Stefan Seyfried QA / R&D Team Mobile Devices | "Any ideas, John?" SUSE LINUX Products GmbH, Nürnberg | "Well, surrounding them's out." ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Suspend-devel mailing list Suspend-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/suspend-devel