From: Dave Hansen <dave@linux.vnet.ibm.com>
To: Andrea Arcangeli <aarcange@redhat.com>
Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org,
Michael J Wolf <mjwolf@us.ibm.com>,
Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [RFC][PATCH 0/6] more detailed per-process transparent hugepage statistics
Date: Tue, 01 Feb 2011 12:56:41 -0800 [thread overview]
Message-ID: <1296593801.27022.3920.camel@nimitz> (raw)
In-Reply-To: <20110201203936.GB16981@random.random>
On Tue, 2011-02-01 at 21:39 +0100, Andrea Arcangeli wrote:
> So now the speedup
> from hugepages needs to also offset the cost of the more frequent
> split/collapse events that didn't happen before.
My concern here is the downward slope. I interpret that as saying that
we'll eventually have _zero_ THPs. Plus, the benefits are decreasing
constantly, even though the scanning overhead is fixed (or increasing
even).
> So I guess considering the time is of the order of 2/3 hours and there
> are "only" 88G of memory, speeding up khugepaged is going to be
> beneficial considering how big boost hugepages gives to the guest with
> NPT/EPT and even bigger boost for regular shadow paging, but it also
> depends on guest. In short khugepaged by default is tuned in a way
> that can't run in the way of the CPU.
I guess we could also try and figure out whether the khugepaged CPU
overhead really comes from the scanning or the collapsing operations
themselves. Should be as easy as some oprofiling.
If it really is the scanning, I bet we could be a lot more efficient
with khugepaged as well. In the case of KVM guests, we're going to have
awfully fixed virtual addresses and processes where collapsing can take
place.
It might make sense to just have split_huge_page() stick the vaddr and
the mm in a queue. khugepaged could scan those addresses first instead
of just going after the system as a whole.
For cases where the page got split, but wasn't modified, should we have
a non-copying, non-allocating fastpath to re-merge it?
-- Dave
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom policy in Canada: sign http://dissolvethecrtc.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2011-02-01 20:56 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-02-01 0:33 [RFC][PATCH 0/6] more detailed per-process transparent hugepage statistics Dave Hansen
2011-02-01 0:33 ` [RFC][PATCH 1/6] count transparent hugepage splits Dave Hansen
2011-02-01 9:58 ` Johannes Weiner
2011-02-03 21:22 ` David Rientjes
2011-02-04 21:18 ` Andrea Arcangeli
2011-02-04 21:28 ` Dave Hansen
2011-02-01 0:33 ` [RFC][PATCH 2/6] pagewalk: only split huge pages when necessary Dave Hansen
2011-02-01 10:04 ` Johannes Weiner
2011-02-01 15:03 ` Dave Hansen
2011-02-03 21:22 ` David Rientjes
2011-02-03 21:33 ` Dave Hansen
2011-02-03 21:46 ` David Rientjes
2011-02-04 17:19 ` Dave Hansen
2011-02-04 21:10 ` Andrea Arcangeli
2011-02-01 0:34 ` [RFC][PATCH 3/6] break out smaps_pte_entry() from smaps_pte_range() Dave Hansen
2011-02-01 10:08 ` Johannes Weiner
2011-02-03 21:22 ` David Rientjes
2011-02-03 21:40 ` Dave Hansen
2011-02-01 0:34 ` [RFC][PATCH 4/6] pass pte size argument in to smaps_pte_entry() Dave Hansen
2011-02-01 10:09 ` Johannes Weiner
2011-02-03 21:22 ` David Rientjes
2011-02-01 0:34 ` [RFC][PATCH 5/6] teach smaps_pte_range() about THP pmds Dave Hansen
2011-02-01 10:11 ` Johannes Weiner
2011-02-01 15:02 ` Dave Hansen
2011-02-01 16:09 ` Andrea Arcangeli
2011-02-03 21:22 ` David Rientjes
2011-02-03 21:34 ` Dave Hansen
2011-02-01 0:34 ` [RFC][PATCH 6/6] have smaps show transparent huge pages Dave Hansen
2011-02-01 10:12 ` Johannes Weiner
2011-02-03 21:22 ` David Rientjes
2011-02-01 15:38 ` [RFC][PATCH 0/6] more detailed per-process transparent hugepage statistics Andrea Arcangeli
2011-02-01 17:15 ` Dave Hansen
2011-02-01 20:39 ` Andrea Arcangeli
2011-02-01 20:56 ` Dave Hansen [this message]
2011-02-02 0:07 ` Andrea Arcangeli
2011-02-08 17:54 ` Dave Hansen
2011-02-08 18:17 ` Andrea Arcangeli
2011-02-03 21:54 ` David Rientjes
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1296593801.27022.3920.camel@nimitz \
--to=dave@linux.vnet.ibm.com \
--cc=aarcange@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mjwolf@us.ibm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).