public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: "SourceForge.net" <noreply@sourceforge.net>
To: noreply@sourceforge.net
Subject: [ kvm-Bugs-1906204 ] AMD NPT causes performance degradation
Date: Wed, 27 Aug 2008 11:41:11 +0000	[thread overview]
Message-ID: <E1KYJOh-0007ft-Kz@h45xhf1.ch3.sourceforge.com> (raw)

Bugs item #1906204, was opened at 2008-03-03 13:45
Message generated for change (Comment added) made by avik
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=893831&aid=1906204&group_id=180599

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Technologov (technologov)
Assigned to: Nobody/Anonymous (nobody)
Summary: AMD NPT causes performance degradation

Initial Comment:
Platform: F7/x64, AMD Barcelona K10, KVM-61.

Guest: Windows XP SP2.

By default, the new Nested Page Tables is enabled, which should accelerate guests.

While it *does* accelerate guests in some areas, particularly guest OS setup time dropped by 20% - which is great, but in other areas I see performance degradation.

For example:
Passmark PerformanceTest v6.1 shows
2D Graphics Marks: 78.6 (without NPT)

2D Graphics Marks: 18.9 (with NPT)

NPT was disabled using:
# rmmod kvm-amd
# modprobe kvm-amd npt=0
# dmesg | tail

and all the graphics feel more sluggish, and way slower.
I have used SDL rendering.

-Alexey, 03.03.2008.

----------------------------------------------------------------------

>Comment By: Avi Kivity (avik)
Date: 2008-08-27 14:41

Message:
Logged In: YES 
user_id=539971
Originator: NO

Make sure your host has the latest BIOS.

----------------------------------------------------------------------

Comment By: Marco Menardi (markit)
Date: 2008-08-26 22:59

Message:
Logged In: YES 
user_id=113918
Originator: NO

Same here, XP has a noticeable slow down. Was faster with my AMD X2 that
is with 4 core Phenom. WIth npt=0 is fast again (as with X2 or better).
Tried with kvm 70...73, debian Sid 64 as host

----------------------------------------------------------------------

Comment By: Anthony Liguori (aliguori)
Date: 2008-03-25 00:03

Message:
Logged In: YES 
user_id=120449
Originator: NO

I am able to reproduce this although the results aren't as extreme.  The
biggest impact is coming from the passmark 2D Graphics Filled Rectangles
test.  Only about 40k exits are occurring while this benchmark is being
run.  It's not impacted by the use of large pages.  Still investigating.


----------------------------------------------------------------------

Comment By: Technologov (technologov)
Date: 2008-03-03 17:25

Message:
Logged In: YES 
user_id=1839746
Originator: YES

Yes, those numbers are always reproducible, even with reloading KVM base
module too.

Just try it for yourself, the benchmark is shareware (free for 30 days) -
but VMs rarely live any longer  anyways :)

link:
http://www.passmark.com/

-Alexey, 03.03.2008.

----------------------------------------------------------------------

Comment By: Joerg Roedel (jroedel)
Date: 2008-03-03 16:40

Message:
Logged In: YES 
user_id=2019182
Originator: NO

Reloading the KVM-AMD module does not disable NPT completly. You have to
reload the KVM base module too. Can you reproduce your numbers if you
reload both and disable NPT?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=893831&aid=1906204&group_id=180599

             reply	other threads:[~2008-08-27 11:41 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-08-27 11:41 SourceForge.net [this message]
  -- strict thread matches above, loose matches on Subject: below --
2008-08-28 23:11 [ kvm-Bugs-1906204 ] AMD NPT causes performance degradation SourceForge.net
2008-08-26 19:59 SourceForge.net
2008-03-03 11:45 SourceForge.net

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=E1KYJOh-0007ft-Kz@h45xhf1.ch3.sourceforge.com \
    --to=noreply@sourceforge.net \
    /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