public inbox for netdev@vger.kernel.org 
 help / color / mirror / Atom feed
From: Jamal Hadi Salim <jhs@mojatatu•com>
To: Jesper Dangaard Brouer <brouer@redhat•com>
Cc: David Miller <davem@davemloft•net>,
	netdev@vger•kernel.org, therbert@google•com,
	hannes@stressinduktion•org, edumazet@google•com,
	jeffrey.t.kirsher@intel•com, rusty@rustcorp•com.au,
	dborkman@redhat•com
Subject: Re: [PATCH 0/2] Get rid of ndo_xmit_flush
Date: Tue, 26 Aug 2014 10:40:28 -0400	[thread overview]
Message-ID: <53FC9C5C.1070200@mojatatu.com> (raw)
In-Reply-To: <20140826121347.0ec7f2ac@redhat.com>

On 08/26/14 06:13, Jesper Dangaard Brouer wrote:
>
> On Tue, 26 Aug 2014 08:28:15 +0200 Jesper Dangaard Brouer <brouer@redhat•com> wrote:
>> On Mon, 25 Aug 2014 16:34:58 -0700 (PDT) David Miller <davem@davemloft•net> wrote:
>>
>>> Given Jesper's performance numbers, it's not the way to go.
>>>
>>> Instead, go with a signalling scheme via new boolean skb->xmit_more.
>>
>> I'll do benchmarking based on this new API proposal today.
>
> While establish an accurate baseline for my measurements.  I'm
> starting to see too much variation in my trafgen measurements.
> Meaning that we unfortunately cannot use it to measure variations on
> the nanosec scale.
>
> I'm measuring the packets per sec via "ifpps", and calculating an
> average over the measurements, via the following oneliner:
>
>   $ ifpps -clod eth5 -t 1000 | awk 'BEGIN{txsum=0; rxsum=0; n=0} /[[:digit:]]/ {txsum+=$11;rxsum+=$3;n++; printf "instant rx:%u tx:%u pps n:%u average: rx:%d tx:%d pps\n", $3, $11, n, rxsum/n, txsum/n }'
>
> Below is measurements done on the *same* kerne:
>   - M1: instant tx:1572766 pps n:215 average: tx:1573360 pps (reboot#1)
>   - M2: instant tx:1561930 pps n:173 average: tx:1557064 pps (reboot#2)
>   - M3: instant tx:1562088 pps n:300 average: tx:1559150 pps (reboot#2)
>   - M4: instant tx:1564404 pps n:120 average: tx:1564948 pps (reboot#3)
>
>   M1->M2: +6.65ns
>   M1->M3: +5.79ns
>   M1->M4: +3.42ns
>   M3->M4: -2.38ns
>
> I cannot explain the variations, but some options could be
>   1) how well the SKB is cache-hot cached via kmem_cache
>   2) other interrups on CPU#0 could disturb us
>   3) interactions with scheduler
>   4) interactions with transparent hugepages
>   5) CPU "turbostat" interactions
>

The clock source used in your system may be an issue as well.
I am not sure if that matters these days but it used to be.
I think back then you could pick ACPI, Jiffies etc to use
as a qdisc clock source.

cheers,
jamal

  parent reply	other threads:[~2014-08-26 14:40 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-08-25 23:34 [PATCH 0/2] Get rid of ndo_xmit_flush David Miller
2014-08-26  6:28 ` Jesper Dangaard Brouer
2014-08-26 10:13   ` Jesper Dangaard Brouer
2014-08-26 12:52     ` Jesper Dangaard Brouer
2014-08-26 16:43       ` Alexander Duyck
2014-08-27  7:48         ` Jesper Dangaard Brouer
2014-08-27  8:37           ` Jesper Dangaard Brouer
2014-08-26 14:40     ` Jamal Hadi Salim [this message]
2014-09-01  0:37     ` Rusty Russell
2014-08-27 12:19 ` Jesper Dangaard Brouer
2014-08-27 20:43   ` David Miller
2014-08-27 12:31 ` Hannes Frederic Sowa
2014-08-27 13:23   ` Eric Dumazet
2014-08-27 13:56     ` Jesper Dangaard Brouer
2014-08-27 14:09       ` Eric Dumazet
2014-08-27 20:48       ` David Miller
2014-08-27 20:46     ` David Miller
2014-08-27 20:45   ` David Miller
2014-08-28  1:42     ` Hannes Frederic Sowa
2014-08-30  3:22       ` David Miller
2014-08-30 10:23         ` Jesper Dangaard Brouer
2014-09-01 20:05         ` Hannes Frederic Sowa
2014-09-01 21:56           ` David Miller
2014-09-01 22:31             ` Hannes Frederic Sowa
2014-09-01 22:35               ` David Miller
2014-08-27 18:28 ` Cong Wang
2014-08-27 19:31   ` Tom Herbert
2014-08-27 20:53     ` David Miller
2014-08-27 20:51   ` David Miller

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=53FC9C5C.1070200@mojatatu.com \
    --to=jhs@mojatatu$(echo .)com \
    --cc=brouer@redhat$(echo .)com \
    --cc=davem@davemloft$(echo .)net \
    --cc=dborkman@redhat$(echo .)com \
    --cc=edumazet@google$(echo .)com \
    --cc=hannes@stressinduktion$(echo .)org \
    --cc=jeffrey.t.kirsher@intel$(echo .)com \
    --cc=netdev@vger$(echo .)kernel.org \
    --cc=rusty@rustcorp$(echo .)com.au \
    --cc=therbert@google$(echo .)com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html
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