From: Simon Jouet <Simon.Jouet@glasgow.ac.uk>
To: Bruce Richardson <bruce.richardson@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] Performance issue with uio_pci_generic driver
Date: Fri, 18 Mar 2016 16:43:09 +0000 [thread overview]
Message-ID: <8379404C2611CB4BB6D4166627C18EA916BB69D4@CMS08-02.campus.gla.ac.uk> (raw)
In-Reply-To: <20160318100957.GB4848@bricha3-MOBL3>
If it’s of any use for debugging I would be willing to setup teamviewer or something like that for a person to remote desktop in a VM and have an SSH access to both servers (all of this is being firewalls that are out of my control so I can’t give direct SSH access). The two servers are for experimentation purposes only so there is nothing critical running.
From: Bruce Richardson<mailto:bruce.richardson@intel.com>
Sent: 18 March 2016 03:11
To: Simon Jouet<mailto:Simon.Jouet@glasgow.ac.uk>
Cc: dev@dpdk.org<mailto:dev@dpdk.org>
Subject: Re: [dpdk-dev] Performance issue with uio_pci_generic driver
On Wed, Mar 16, 2016 at 11:59:34PM +0000, Simon Jouet wrote:
> Hi everyone,
>
> First off I would like to thanks tmonjalo, Harry Van Harren and Bruce Richardson for the input they gave while I was trying to figure out the issue and pushing me to report the problem here 😊
>
> Okay, so I was trying out some basic sanity benchmarks with DPDK before doing anything more complicated and surprisingly I was getting lower than gigabit speed for minimum packet size running l2fwd (or l3fwd for that matter).
>
> The setup is very simple I’ve got two machine with Intel x710 quad port NICs one is running DPDK l2fwd and the other is running MoonGen for the performance benchmark.
>
> After much debugging and trying to modify parameters one by one, giving up after nothing worked and setting up ovs-dpdk I noticed from the ovs documentation that the kernel module to load were uio and igb_uio while I was previous loading uio_pci_generic as mentioned in the DPDK getting started guide. I simply changed the kernel module and l2fwd went from 700Mbps to 10G line-rate. Bruce said that shouldn’t be the case and the performance should be similar regardless of the driver loaded ...
>
> Here is the full log of the experiment, if you’re interested:
> https://gist.github.com/simon-jouet/178e1d302afef5c6a642
>
> Best regards,
> Simon
>
Thanks Simon for reporting this here!
It seems very strange to me that the particular module being used for uio has
such an impact on performance. Has anyone else seen anything similar?
/Bruce
prev parent reply other threads:[~2016-03-18 16:43 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-03-16 23:59 Simon Jouet
2016-03-18 10:09 ` Bruce Richardson
2016-03-18 16:43 ` Simon Jouet [this message]
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=8379404C2611CB4BB6D4166627C18EA916BB69D4@CMS08-02.campus.gla.ac.uk \
--to=simon.jouet@glasgow.ac.uk \
--cc=bruce.richardson@intel.com \
--cc=dev@dpdk.org \
/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).