[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Tue, 7 Jul 2009 13:07:44 -0700 (PDT)
From: Dan Magenheimer <dan.magenheimer@...cle.com>
To: Rik van Riel <riel@...hat.com>
Cc: linux-kernel@...r.kernel.org, npiggin@...e.de, akpm@...l.org,
jeremy@...p.org, xen-devel@...ts.xensource.com,
tmem-devel@....oracle.com, alan@...rguk.ukuu.org.uk,
linux-mm@...ck.org, kurt.hackel@...cle.com,
Rusty Russell <rusty@...tcorp.com.au>,
dave.mccracken@...cle.com, Marcelo Tosatti <mtosatti@...hat.com>,
sunil.mushran@...cle.com, Avi Kivity <avi@...hat.com>,
Schwidefsky <schwidefsky@...ibm.com>, chris.mason@...cle.com,
Balbir Singh <balbir@...ux.vnet.ibm.com>
Subject: RE: [RFC PATCH 1/4] (Take 2): tmem: Core API between kernel and tmem
> From: Rik van Riel [mailto:riel@...hat.com]
> Subject: Re: [RFC PATCH 1/4] (Take 2): tmem: Core API between
>
> Dan Magenheimer wrote:
> > Tmem [PATCH 1/4] (Take 2): Core API between kernel and tmem
>
> I like the cleanup of your patch series.
Thanks much, but credit goes to Jeremy for suggesting this
very clean tmem_ops interface.
> However, what remains is a fair bit of code.
Yes, though much of the LOC is for clean layering and
readability. (Nearly half of the patch is now comments.)
> It would be good to have performance numbers before
> deciding whether or not to merge all this code.
On one benchmark that I will be presenting at Linux Symposium
(8 dual-VCPU guests with 384MB of initial memory and doing
self-ballooning to constrain memory, each guest compiling
Linux continually; quad-core-dual-thread Nehalem processor
with 4GB physical RAM) I am seeing savings of ~300 IO/sec
at an approximate cost of 0.1%-0.2% of one CPU. But
I admit much more benchmarking needs to be done.
Thanks,
Dan
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists