From: Maayan Kashani <mkashani@nvidia.com>
To: Kamil Vojanec <vojanec@cesnet.cz>, "dev@dpdk.org" <dev@dpdk.org>
Cc: Jan Viktorin <viktorin@cesnet.cz>
Subject: RE: ConnectX-7 performance anomaly
Date: Thu, 10 Apr 2025 09:48:56 +0000 [thread overview]
Message-ID: <PH7PR12MB70183E112D4800F12ED89BFEB2B72@PH7PR12MB7018.namprd12.prod.outlook.com> (raw)
In-Reply-To: <f2317379-ee61-42b2-bee6-10efaa91e2ca@cesnet.cz>
Hi, Kamil,
Thanks for reaching out.
After short sync with our performance team,
I can tell we did not observe such anomality and it
seems to be some local server issue/configurations.
Regards,
Maayan Kashani
> -----Original Message-----
> From: Kamil Vojanec <vojanec@cesnet.cz>
> Sent: Thursday, 3 April 2025 14:28
> To: dev@dpdk.org
> Cc: Jan Viktorin <viktorin@cesnet.cz>
> Subject: ConnectX-7 performance anomaly
>
> Hello,
>
> while benchmarking the Nvidia ConnectX-7 400Gbe NIC, I encountered some
> interesting behavior: when processing short packets (72 B), the maximum
> packet rate I can achieve is 230 Mpps. However, when I increase the
> packet size to 512 B briefly and then bring the packet size down to 72 B
> again, the maximum throughput is now 260 Mpps and seems to stabilize at
> that value. According to the Nvidia performance report
> (https://fast.dpdk.org/doc/perf/DPDK_24_07_NVIDIA_NIC_performance_re
> port.pdf),
> the maximum achievable packet rate of ConnectX-7 NICs is 266 Mpps and
> that is what I am trying to reproduce. I have observed this behavior
> with dpdk-l3fwd as well as with my own DPDK application.
>
> My setup is as follows:
>
> CPU: AMD EPYC 9554P 64-Core Processor
>
> Application commandline:
>
> sudo dpdk_example_l3fwd -a0000:05:00.0 -l0-32 --main-lcore=32 -- -p0x1
> --config
> "(0,0,0),(0,1,1),(0,2,2),(0,3,3),(0,4,4),(0,5,5),(0,6,6),(0,7,7),(0,8,8),(0,9,9),(0,1
> 0,10),(0,11,11),(0,12,12),(0,13,13),(0,14,14),(0,15,15),(0,16,16),(0,17,17),(0,1
> 8,18),(0,19,19),(0,20,20),(0,21,21),(0,22,22),(0,23,23),(0,24,24),(0,25,25),(0,2
> 6,26),(0,27,27),(0,28,28),(0,29,29),(0,30,30),(0,31,31)"
> --eth-dest=0,00:10:94:00:00:02
>
> Generator: Spirent 400G generator
>
> Would anyone be able to share some insight as to what might be happening
> here?
>
> Kamil Vojanec
> CESNET z.s.p.o.
prev parent reply other threads:[~2025-04-10 9:49 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-04-03 11:27 Kamil Vojanec
2025-04-10 9:48 ` Maayan Kashani [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=PH7PR12MB70183E112D4800F12ED89BFEB2B72@PH7PR12MB7018.namprd12.prod.outlook.com \
--to=mkashani@nvidia.com \
--cc=dev@dpdk.org \
--cc=viktorin@cesnet.cz \
--cc=vojanec@cesnet.cz \
/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).