DPDK patches and discussions
 help / color / mirror / Atom feed
* Re: [dpdk-dev] Why pktgen-dpdk will receive more packet than sent
@ 2015-06-30  3:15 Wiles, Keith
  0 siblings, 0 replies; 2+ messages in thread
From: Wiles, Keith @ 2015-06-30  3:15 UTC (permalink / raw)
  To: Jiang, Yunhong, dev



On 6/29/15, 9:01 PM, "dev on behalf of Jiang, Yunhong"
<dev-bounces@dpdk.org on behalf of yunhong.jiang@intel.com> wrote:

>Hi, all
>	I'm trying to test the l2fwd throughput using
>https://github.com/pktgen/Pktgen-DPDK/blob/master/dpdk/examples/pktgen/scr
>ipts/rfc2544.lua .  However, the pktgen will receive more packets than it
>sent out like " Total sent 14786446 recv 14786597, delta 151". I'm really
>confused why this will happen. I have the pktgen machine and l2fwd
>machine connected to each other directly and bind the igb_uio to the NIC
>devices.
>
>Pktgen> *** Test packet size 64 at rate 100%
>Total sent 133856261 recv 131517412, delta -2338849
>*** Test packet size 64 at rate 50%
>Total sent 40208555 recv 40207420, delta -1135
>*** Test packet size 64 at rate 25%
>PMD: ixgbe_set_rx_function(): Vector rx enabled, please make sure RX
>burst size no less than 32.
>Total sent 14786446 recv 14786597, delta 151
>*** Test packet size 64 at rate 37%
>PMD: ixgbe_set_rx_function(): Vector rx enabled, please make sure RX
>burst size no less than 32.
>Total sent 26824256 recv 26823587, delta -669

The number of packets sent/recv are read from the NIC counters and Pktgen
does not count the packets in anyway.

The script could also be in error. I was just looking at the script and
maybe the timing from stoping the transmit and reading the counters is the
issue. You could try putting a delay before the counter read to make sure
all of the packets have been sent/recv along with the reading of the
packets.

I do not have a setup to test it at this time.

Regards,
++Keith
>
>Any hints will be really helpful, thanks!
>
>Thanks
>--jyh

^ permalink raw reply	[flat|nested] 2+ messages in thread

* [dpdk-dev] Why pktgen-dpdk will receive more packet than sent
@ 2015-06-30  2:01 Jiang, Yunhong
  0 siblings, 0 replies; 2+ messages in thread
From: Jiang, Yunhong @ 2015-06-30  2:01 UTC (permalink / raw)
  To: dev

Hi, all
	I'm trying to test the l2fwd throughput using https://github.com/pktgen/Pktgen-DPDK/blob/master/dpdk/examples/pktgen/scripts/rfc2544.lua .  However, the pktgen will receive more packets than it sent out like " Total sent 14786446 recv 14786597, delta 151". I'm really confused why this will happen. I have the pktgen machine and l2fwd machine connected to each other directly and bind the igb_uio to the NIC devices.

Pktgen> *** Test packet size 64 at rate 100%
Total sent 133856261 recv 131517412, delta -2338849
*** Test packet size 64 at rate 50%
Total sent 40208555 recv 40207420, delta -1135
*** Test packet size 64 at rate 25%
PMD: ixgbe_set_rx_function(): Vector rx enabled, please make sure RX burst size no less than 32.
Total sent 14786446 recv 14786597, delta 151
*** Test packet size 64 at rate 37%
PMD: ixgbe_set_rx_function(): Vector rx enabled, please make sure RX burst size no less than 32.
Total sent 26824256 recv 26823587, delta -669

Any hints will be really helpful, thanks!

Thanks
--jyh

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2015-06-30  3:15 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-06-30  3:15 [dpdk-dev] Why pktgen-dpdk will receive more packet than sent Wiles, Keith
  -- strict thread matches above, loose matches on Subject: below --
2015-06-30  2:01 Jiang, Yunhong

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).