Found another system (Core(TM) i7-2760QM on Ubuntu 12.04) and got 811674 pps and data rate of 59.079 MiB/t
Wonder if it's possible to do better. Thanks again. On Monday, June 8, 2015 at 5:18:43 PM UTC-5, Daniel Borkmann wrote: > > On 06/09/2015 12:11 AM, Perf Tech wrote: > > Thanks for catching the problem. Here is the output you asked for. > > > > # cat /proc/net/packet > > sk RefCnt Type Proto Iface R Rmem User Inode > > ffff8800d8a2d800 3 3 0003 4 1 0 0 10933 > > Ok, if you've run that while running trafgen, that it's the trafgen > packet socket. Otherwise, you could also try to kill that application, > as that might also consume cycles (in case your kernel doesn't support > direct xmit mode). Other than that and other than this negotiation > issue, you should also dig a bit into the link from that blog, etc. > > > It's strange that both ports can support T1000/Full according to the > output > > of "ethtool eth0", "ethtool eth1", but they negotiated to be 100Mbps. > Need > > to figure out why. > > > > Thanks again! > > > > > > On Monday, June 8, 2015 at 5:08:26 PM UTC-5, Daniel Borkmann wrote: > >> > >> On 06/09/2015 12:01 AM, Perf Tech wrote: > >>> Thanks for the suggestion. Here is the output of ifpps. It shows the > >>> kernel. I didn't do any tunning on this system. > >> > >> Ok, I thought you were talking about 1G, but your link is just 100Mbit. > >> Do you have anything running in the background of your system that is > >> listening on a packet socket? > >> > >> cat /proc/net/packet > >> > >>> 3.16.0-23-generic, x86_64, eth0 (e1000e 100Mbit/s link:yes), > >> t=1000ms, > >>> cpus=4/ > >>> 4 > >>> rx: 0.000 MiB/t 0 pkts/t 0 drops/t > >>> 0 err > >>> tx: 9.446 MiB/t 128636 pkts/t 0 drops/t > >>> 0 err > >>> ors/t > >>> rx: 0.000 MiB 0 pkts 0 drops > >>> 0 err > >>> ortx: 883.624 MiB 12033079 pkts 0 drops > >>> 0 err > >>> ors > >>> sys: 726 cs/t 143 procs 5 running > >>> 0 iow > >>> ait > >>> mem: 3833M total 568M used 348M active > >>> 121M ina > >>> ctswap: 3971M total 0M used 0M cached > >>> > >>> cpu3 +: 6.9% usr/t 93.1% sys/t 0.0% idl/t > >>> 0.0% iow > >>> /tcpu1 |: 5.0% usr/t 95.0% sys/t 0.0% idl/t > >>> 0.0% iow > >>> /tcpu0 |: 5.3% usr/t 94.7% sys/t 0.0% idl/t > >>> 0.0% iow > >>> /tcpu2 -: 5.1% usr/t 94.9% sys/t 0.0% idl/t > >>> 0.0% iow > >>> /tavg: 5.4% 94.6% 0.0% > >> 0.0% > >>> > >>> cpu0 +: 4012 irqs/t 4012 sirq rx/t > >> 4012 > >>> sirq > >>> tcpu3 |: 0 irqs/t 1 sirq rx/t > >> 0 > >>> sir > >>> > >>> > >>> On Monday, June 8, 2015 at 4:55:07 PM UTC-5, Daniel Borkmann wrote: > >>>> > >>>> On 06/08/2015 11:46 PM, Daniel Borkmann wrote: > >>>>> On 06/06/2015 01:05 AM, Perf Tech wrote: > >>>>>> I have a PC with 3 NICs (eth0 is Intel 82579LM NIC, eth1 is > 82574L), > >>>> OS is > >>>>>> ubuntu 12.04 server. > >>>>>> I connected eth0 and eth1 with a ethernet cable and both ports are > >> up. > >>>>>> > >>>>>> Trying to find out how fast trafgen(0.5.9-rc4) can send packets on > >> this > >>>>>> system. > >>>>>> > >>>>>> # ./trafgen --dev eth0 --conf tmp2.txt --cpp > >>>>>> 4 packets to schedule > >>>>>> 292 bytes in total > >>>>>> Running! Hang up with ^C! > >>>>>> > >>>>>> > >>>>>> I am monitoring "ifconfig eth0" with a monitoring program that I > have > >>>> used > >>>> > >>>> There's a tool in the netsniff-ng repo called ifpps, you might also > >> want > >>>> to > >>>> check that out. > >>>> > >>>>>> many years, the issue is, trafgen is only sending > >>>>>> pcap at a rate 129000 packets/sec with data rate of 79Mbps. I had > >>>>>> expected it to be a long higher. > >>>>> > >>>>> What kernel and NIC driver do you use? Did you do any particular > >> tuning > >>>>> settings? > >>>>> > >>>>> E.g. a starting point could be: > >>>>> > >>>>> > >>>> > >> > http://netoptimizer.blogspot.ch/2014/09/network-setup-for-accurate-nanosec.html > > >>>>> > >>>>> Also, if your config uses checksumming, it's done in trafgen in > >>>> software, > >>>>> so if you want to test a simple minimal config, you could use a > static > >>>> buffer > >>>>> as a start. > >>>>> > >>>>>> > >>>>>> Did I run the command wrong? > >>>>>> > >>>>>> Thanks > >>>>>> > >>>>> > >>>> > >>>> > >>> > >> > >> > > > > -- You received this message because you are subscribed to the Google Groups "netsniff-ng" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
