DPDK patches and discussions
 help / color / mirror / Atom feed
From: mukawa@igel.co.jp
To: dev@dpdk.org
Cc: nakajima.yoshihiro@lab.ntt.co.jp, masutani.hitoshi@lab.ntt.co.jp
Subject: [dpdk-dev] [RFC] PMD for performance measurement
Date: Fri, 19 Sep 2014 21:27:38 +0900	[thread overview]
Message-ID: <1411129659-7132-1-git-send-email-mukawa@igel.co.jp> (raw)
In-Reply-To: <mukawa@igel.co.jp>

From: Tetsuya Mukawa <mukawa@igel.co.jp>

Hi,

I want to measure throughputs like following cases.

- path connected by RING PMDs
- path connected by librte_vhost and virtio-net PMD
- path connected by MEMNIC PMDs
- .....

Is there anyone want to do same thing?

Anyway, I guess those throughputs may be too high for some devices like ixia.
But it's a bit pain to write or fix applications just for measuring.

I guess a PMD like '/dev/null' and testpmd application will help.
This patch set is RFC of a PMD like '/dev/null'.
Please see the first commit of this patch set.


Here is a my plan to use this PMD

+-------------------------------+
|           testpmd1            |
+-------------+------+----------+
| Target PMD1 |      | null PMD |
+---++--------+      +----------+
    ||
    || Target path
    ||
+---++--------+      +----------+
| Target PMD2 |      | null PMD |
+-------------+------+----------+
|           testpmd2            |
+-------------------------------+

The testpmd1 or testpmd2 will start with "start tx_first". It causes huge
transfers.

The result is not thuroughput of PMD1 or PMD2, but throughput
between PMD1 and PMD2. But I guess it's enough to know rough thuroughput.
Also it's nice for measuing that the same environment can be used.

Any suggestions or comments?

Thanks,
Tetsuya

--
Tetsuya Mukawa (1):
  librte_pmd_null: Add null PMD

 config/common_bsdapp               |   5 +
 config/common_linuxapp             |   5 +
 lib/Makefile                       |   1 +
 lib/librte_pmd_null/Makefile       |  58 +++++
 lib/librte_pmd_null/rte_eth_null.c | 474 +++++++++++++++++++++++++++++++++++++
 5 files changed, 543 insertions(+)
 create mode 100644 lib/librte_pmd_null/Makefile
 create mode 100644 lib/librte_pmd_null/rte_eth_null.c

-- 
1.9.1

             reply	other threads:[~2014-09-19 12:23 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-09-19 12:27 mukawa [this message]
2014-09-19 12:27 ` [dpdk-dev] [RFC] librte_pmd_null: Add null PMD mukawa
2014-09-22  7:46   ` Ivan Boule
2014-09-22  8:07     ` Tetsuya Mukawa

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=1411129659-7132-1-git-send-email-mukawa@igel.co.jp \
    --to=mukawa@igel.co.jp \
    --cc=dev@dpdk.org \
    --cc=masutani.hitoshi@lab.ntt.co.jp \
    --cc=nakajima.yoshihiro@lab.ntt.co.jp \
    /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).