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

Re: [Xen-devel] [PATCH 1 of 3] libxl: take node distances into account during NUMA placement

On Fri, 2012-10-19 at 11:39 +0100, Stefano Stabellini wrote:
> On Fri, 19 Oct 2012, Ian Jackson wrote:
> > Dario Faggioli writes ("Re: [Xen-devel] [PATCH 1 of 3] libxl: take node 
> > distances into account during NUMA placement"):
> > > It is, with N being the number of nodes, which we discussed thoroughly
> > > already a couple of months ago, and reached consensus on the fact that N
> > > will stay less than 8 for the next 5 (but probably even more) years. :-)
> > 
> > No, I don't think we did reach consensus about that.  It was asserted
> > but I dissented.  I don't think this is a reasonable assumption.
> When Dario speaks about consensus in terms of hardware that is going to
> reach the market, he really means what Intel and AMD have told us.

That being said, I really wanted to avoid re-starting that discussion
so, while trying to summarize it as quickly as possible, I've no problem
admitting that "reach consensus" could not be the perfect choice of
words. My bad.

The whole point is that the solution we have (and that I'm trying to
keep up improving with these patches) is both viable with current and
near future hardware and no harms with any crazy unexpectable
breakthrough in CPU design --thanks to the safety catch IanJ suggested. 

Moreover, it is easily amendable, for example taking more advantage of
cpupools (with witch the placement algorithm is already integrated up to
some extent), in case the latter happens.
This is so true that, as I already said, I've no problem even starting
to think about how to put it together. Maybe not from tomorrow (I'm
quite busy with other stuff :-D), but definitely before 4.3, if we think
it's something we couldn't live without.

> Sorry but your opinion doesn't count that much in the matter of cpu
> architecture being produces in the near future, unless you have already
> started a secret Californian cpu startup, that nobody knows about yet
> ;-)


<<This happens because I choose it to happen!>> (Raistlin Majere)
Dario Faggioli, Ph.D, http://retis.sssup.it/people/faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

Attachment: signature.asc
Description: This is a digitally signed message part

Xen-devel mailing list



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