[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20181221164324.5bd7b5b9@redhat.com>
Date: Fri, 21 Dec 2018 16:43:24 +0100
From: Jesper Dangaard Brouer <brouer@...hat.com>
To: Saeed Mahameed <saeedm@...lanox.com>
Cc: "David S. Miller" <davem@...emloft.net>, netdev@...r.kernel.org,
Tariq Toukan <tariqt@...lanox.com>,
Tal Gilboa <talgi@...lanox.com>,
Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
Pawel Staszewski <pstaszewski@...are.pl>, brouer@...hat.com,
Toke Høiland-Jørgensen <toke@...e.dk>
Subject: Re: [pull request][net-next 0/9] Mellanox, mlx5 XDP TX 100 Million
packet per second
On Thu, 20 Dec 2018 23:06:38 -0800
Saeed Mahameed <saeedm@...lanox.com> wrote:
> XDP_TX:
> We see a huge gain on single port ConnectX-5, and reach the 100 Mpps
> milestone.
> * Single-port HCA:
> Before: 70 Mpps
> After: 100 Mpps (+42.8%)
Great, this confirm that we were hitting a PCIe limit in our XDP-paper.
See graph[1] leveling off at 70Mpps for XDP_TX.
[1] https://github.com/xdp-project/xdp-project/blob/master/conference/LinuxPlumbers2018/images/redirect-test.pdf
--
Best regards,
Jesper Dangaard Brouer
MSc.CS, Principal Kernel Engineer at Red Hat
LinkedIn: http://www.linkedin.com/in/brouer
Powered by blists - more mailing lists