VM entry/exit time

VM entry/exit time

[Omar Khan]

hi,

   I am measuring the VM enty/exit time on the following two machines by
recording the time it takes to make n number of vmcalls :

1. Lenovo Think Centre, Intel Pentium D, 2 Cores, 3GHz with 3G of memory
2. Dell Inspiron 6400 , Intel Core Duo, 2 Cores, 2 GHz with 1G of memory

Surprisingly it takes almost 3 times longer on the Lenovo to make a vmcall.
Anyone have any idea what might be causing this? Running kvm on suse10.2 on both
the systems.

Thanks 

Omar Khan 


-------------------------------------------------------------------------
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/

Re: VM entry/exit time

[Avi Kivity]

Omar Khan wrote:
> hi,
>
>    I am measuring the VM enty/exit time on the following two machines by
> recording the time it takes to make n number of vmcalls :
>
> 1. Lenovo Think Centre, Intel Pentium D, 2 Cores, 3GHz with 3G of memory
> 2. Dell Inspiron 6400 , Intel Core Duo, 2 Cores, 2 GHz with 1G of memory
>
> Surprisingly it takes almost 3 times longer on the Lenovo to make a vmcall.
> Anyone have any idea what might be causing this? Running kvm on suse10.2 on both
> the systems.
>
>   

It's not surprising at all.  VT is improving at a rapid pace.  I suspect 
that on a Core 2 you will see faster speeds, and more improvements are 
promised with newer processors.

Note that recent kvm versions have improved vmexit latency independently 
of the processor, and this is continuing too.


-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
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/