[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] scaling problem with writable pagetables



Keir Fraser wrote:

On 15 Feb 2006, at 18:49, Andrew Theurer wrote:

Below are per-function ratios of CPU time rev8830/rev8830-lock-reduction (derived from oprofile diffs)

SDET:

9.84/1   restore_all_guest

Kind of odd, since that function contains no locking. Perhaps VCPUs are blocked so long that, by the time they get the CPU there is an event pending and the hypercall gets preempted?
I was a bit surprised at that, too.

What do the perfctr numbers look like for #hypercalls and #exceptions? Also worth adding one to __hypercall_create_continuation as that'll count #preemptions.
I have not taken perfctr numbers yet, but I will today.  Thanks,

-Andrew


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.