On Mon, 2007-06-25 at 18:25 -0400, Avi Kivity wrote: > Gregory Haskins wrote: > > On Mon, 2007-06-25 at 18:08 -0400, Avi Kivity wrote: > > > >> There's a slight issue that Windows with the ACPI HAL > >> is dog slow when virtulalized, but maybe we can find another HAL that > >> supports ACPI but doesn't use the APIC. > >> > > > > Is this in reference to the frequent TPR updates, or is there something > > else that is also slow with ACPI enabled? > > > > > > Just the good old tpr.
Ah, yes. I actually find the later versions of KVM to perform quite well even with XP+ACPI. I think its all the work that went into the VMX light-exits, mostly. Things seem to run even better with the in-kernel APIC patch enabled, but admittedly I haven't run any benchmarks on it to tell for sure. This is all "seat of the pants" ;). I am not exactly sure why the in-kernel APIC would/could make a difference since TPRs are light-exits in either case IIUC... I am very interested to see what kind of gains we can get when TPR shadowing is implemented properly (my attempt wasn't even fully functional if you recall). I seem to recall reading somewhere that TPR-shadow features can work even with 32 bit OSes (I guess MOV-TO-CR8 works in 32-bit mode?) but I don't know if XP would use it or the MMIO path. Im guessing MMIO, but perhaps MS will patch it since they are interested in VMM technology now. Does AMD support a similar notion of TPR-shadow? -Greg ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ kvm-devel mailing list kvm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/kvm-devel
