From: Stephen Hemminger <stephen@networkplumber.org>
To: 曾懷恩 <the@csie.io>
Cc: users@dpdk.org
Subject: Re: [dpdk-users] DPDK and isolcpus
Date: Fri, 26 Jul 2019 07:56:09 -0700 [thread overview]
Message-ID: <20190726075609.4730000a@hermes.lan> (raw)
In-Reply-To: <202CB4D5-3F29-44C1-AC0F-BD5AD8C91782@csie.io>
On Fri, 26 Jul 2019 22:50:47 +0800
曾懷恩 <the@csie.io> wrote:
> Hi Filip,
>
> Do you use Mellanox NIC?
>
> In my experience, rte_lcore_count only counts those lcores that are not isolated when using mlx5 driver.
>
> And Intel ixgbe counts lcores correctly even I use isolcpus flag
>
> BR,
>
> Sent from my iPhone
>
> > From: Filip Janiszewski <contact@filipjaniszewski.com>
> > To: "users@dpdk.org" <users@dpdk.org>
> > Subject: [dpdk-users] DPDK and isolcpus
> > Message-ID:
> > <e98abf2a-b1cf-5fee-7929-778f70390d1d@filipjaniszewski.com>
> > Content-Type: text/plain; charset=iso-8859-15
> >
> > Hi,
> >
> > I've configured a bunch of my box cores with isolcpus and nohz_full in
> > the attempt to squeeze a little more performance out of them, but,
> > apparently I can't use those core in DPDK anymore as it seems that
> > rte_lcore_count reports only core which are not isolated, also I can't
> > launch any thread (rte_eal_remote_launch) on those cores.
> >
> > In my understanding
> > (http://doc.dpdk.org/spp-18.02/setup/performance_opt.html) I should be
> > able to "Use the isolcpus Linux kernel parameter to isolate them from
> > Linux scheduler to reduce context switches. It prevents workloads of
> > other processes than DPDK running on reserved cores with isolcpus
> > parameter." but can't make it work.
> >
> > Which is the correct way to bind DPDK threads to isolated cpus?
> >
> > Thanks
> >
> > --
> > BR, Filip
> > +48 666 369 823
> >
> >
> > End of users Digest, Vol 196, Issue 5
> > *************************************
The default behavior of DPDK is to use all CPU's it finds.
To use isolated CPU's you need to pass the cpu's to use on the command line.
dpdk-XXX -l 4-7 ...
next prev parent reply other threads:[~2019-07-26 14:56 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <mailman.1.1564135202.18507.users@dpdk.org>
2019-07-26 14:50 ` 曾懷恩
2019-07-26 14:56 ` Stephen Hemminger [this message]
2019-07-26 14:56 曾懷恩
-- strict thread matches above, loose matches on Subject: below --
2019-07-26 6:17 Filip Janiszewski
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=20190726075609.4730000a@hermes.lan \
--to=stephen@networkplumber.org \
--cc=the@csie.io \
--cc=users@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).