|  |  | 
  
    |  |  | 
 
  |   |  | 
  
    |  |  | 
  
    |  |  | 
  
    |   xen-users
Re: [Xen-users] Xen 3.1 - 32 vs 64 bit hypervisor 
| On Fri, Jun 08, 2007 at 03:36:02PM +0100, Mark Williamson wrote:
> > I've been experimenting with Xen 3.1.
> >
> > My requirements are to run lots 32 bit paravirt domains on large
> > memory hosts (8GB or more).
> >
> > Previously I've used a 32 bit Xen with PAE hypervisor which works
> > well.  However I seen from the announcement that you can use a 64 bit
> > hypervisor and still run 32 bit PAE domains.  I've tried this and it
> > works as advertised!
> 
> Awesome :-)
> 
> > What I'm wondering is, will it be best for me to run a 32 bit PAE Xen
> > hypervisor or a 64 bit hypervisor?  What is the tradeoff between the
> > two? Which is likely to be more efficient and which more stable?
> 
> Stability-wise, 32-on-64 is a new feature so it will have received less real 
> world abuse testing.  There may be teething troubles.  Performance-wise...  
> Again, the combination is less tested, so it's not inconceivable that there 
> will be teething troubles.
> 
> That said, in principle it ought to be alright.  32-on-64 even fully supports 
> glibc's TLS, which even 32-on-32 doesn't (this is because of slightly arcane 
> implementation details in x86.  Really.  The mind boggles.).
I won't ask as I don't need my brain exploded on Friday afternoon ;-)
> If you're running a 64-bit Xen, I think you also have the potential to run 
> 64-bit paravirt guests, and 64-bit HVM guests in addition to all the stuff 
> you could run on 32-bit Xen.
Yes that is certainly of future interest.
> At some point it should be possible to run a 32-bit dom0, but I'm not sure if 
> that's allowed yet.
Well that is how I tested it so I guess it is possible ;-)
Thanks for your useful comments!
-- 
Nick Craig-Wood <nick@xxxxxxxxxxxxxx> -- http://www.craig-wood.com/nick
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
 | 
 |  | 
  
    |  |  |