Thanks, Stephen. That's the problem.
On Sun, Aug 10, 2014 at 2:11 PM, Stephen Hemminger <
stephen at networkplumber.org> wrote:
> On Sat, 9 Aug 2014 11:59:24 -0400
> Kai Zhang wrote:
>
> > Hi there,
> >
> > I have a problem for multi-queue receiving.
&g
Hi there,
I have a problem for multi-queue receiving.
For a specific application, I am not using the lcore abstractions in DPDK.
I launch rx and tx threads seperately. For one port with N queues, I
launched N rx threads and N tx threads, with each thread in charge of
receiving or sending only one
Thanks!
I have found the problem. It is that I used clock_gettime() to measure the
latency. With rte_rdtsc(), the round trip latency is measured to be less
than 10 microseconds.
Thanks very much,
Kai
On Thu, Jul 24, 2014 at 8:48 AM, Wodkowski, PawelX <
pawelx.wodkowski at intel.com> wrote:
> R
Hello,
I am trying to develop a low-latency application, and I measured the round
trip latency with DPDK. However I got an average of 650~720 microseconds
round-trip latency with Intel 82599 10Gbps NIC.
The experiment method is as follows. 2 machines (A and B) are connected
back-to-back. Machine
Hi Bruce,
Thanks for your help. I have implemented rx and tx in the same program to
measure round trip latency.
2 machines (A and B) are connected back-to-back with Intel 82599 NIC.
Machine A embeds a time stamp in the packet and sends to B, B (use testpmd
or l2fwd) forwards packets back to A imm
Hello,
I have been trying to measure the round trip latency with two machines,
A->B->A. I modified l2fwd as two new apps: rx and tx, where tx sends
packets and embeds time stamp, rx receives packets and calculate the round
trip latency.
The source code are as follows:
https://github.com/kay21s/dpd
6 matches
Mail list logo
