freebsd stable - Jun 2006 - Gigabit ethernet very slow.

What's up with my computer, it's only getting 30MB/s?

hostB: nc -4kl port > /dev/null
hostA: nc host port < /dev/zero

hostB: Athlon64 3000, Asus A8R-MVP, FreeBSD 6.1/amd64):
$ sysctl -a net.inet.tcp|grep space
net.inet.tcp.sendspace: 32768
net.inet.tcp.recvspace: 131072
sk0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 6500
        options=8<VLAN_MTU>
        inet6 fe80::215:f2ff:fed9:ae2a%sk0 prefixlen 64 scopeid 0x2
        inet 192.168.1.242 netmask 0xffffff00 broadcast 192.168.1.255
        ether 00:15:f2:d9:ae:2a
        media: Ethernet 1000baseTX <full-duplex> (1000baseTX
<full-duplex,flag0,flag1>)
        status: active
top:
last pid:   730;  load averages:  0.53,  0.41,  0.41
26 processes:  1 running, 25 sleeping
CPU states:  2.6% user,  0.0% nice, 12.4% system, 25.9% interrupt, 59.0% idle
Mem: 11M Active, 7832K Inact, 23M Wired, 10M Buf, 189M Free
Swap: 483M Total, 483M Free

  PID USERNAME  THR PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
  557 root        1  98    0  3964K   956K select   7:48  9.57% nc

$ vmstat -n 0 5
 procs      memory      page                      faults      cpu
 r b w     avm    fre  flt  re  pi  po  fr  sr   in   sy  cs us sy id
 1 0 0   31368 194048   15   0   0   0  13   0 12705 110013 21101  1 38 61
 1 0 0   31368 194048    0   0   0   0   0   0 9638 84515 14842  1 28 71
 0 0 0   31368 194048    0   0   0   0   0   0 10228 89549 15756  1 31 68
 0 0 0   31368 194048    0   0   0   0   0   0 13055 114963 20048  1 38 60
 0 0 0   31368 194048    0   0   0   0   0   0 9325 81653 14384  0 28 72


---------------------------------------


hostA: P4 3GHz Prescott, Intel SE7210TP1E, FreeBSD 6.1/i386):
$ sysctl -a net.inet.tcp|grep space
net.inet.tcp.sendspace: 65536
net.inet.tcp.recvspace: 65536
em0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 6500
        options=b<RXCSUM,TXCSUM,VLAN_MTU>
        inet6 fe80::204:23ff:feba:3870%em0 prefixlen 64 scopeid 0x1
        inet 192.168.1.2 netmask 0xffffff00 broadcast 192.168.1.255
        ether 00:04:23:ba:38:70
        media: Ethernet 1000baseTX <full-duplex>
        status: active
top:
last pid:  1190;  load averages:  0.10,  0.21,  0.28
50 processes:  1 running, 49 sleeping
CPU states:  0.7% user,  0.0% nice, 11.6% system,  9.0% interrupt, 78.7% idle
Mem: 72M Active, 222M Inact, 147M Wired, 12K Cache, 111M Buf, 552M Free
Swap: 6144M Total, 6144M Free

  PID USERNAME  THR PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
  997 root        1   4    0  1528K   720K sbwait   1:37 15.82% nc

$ vmstat -n 0 5
 procs      memory      page                      faults      cpu
 r b w     avm    fre  flt  re  pi  po  fr  sr   in   sy  cs us sy id
 2 0 0  169608 563404   36   0   0   0  28   0 9149 109489 14087  3 34 64
 0 0 0  169608 563404    9   0   0   0   9   0 8224 92285 11696  2 24 74
 1 0 0  169608 563404   13   0   0   0  13   0 8818 99341 12541  2 28 70
 0 0 0  169608 563404    9   0   0   0   9   0 8645 97585 12303  2 28 70
 0 0 0  169608 563404   13   0   0   0  13   0 7610 85517 10830  1 26 73

On 6/25/06, Sean Bryant <bryants@gmail.com> wrote:
> /dev/zero not exactly the best way to test sending data across the
> network. Especially since you'll be reading a 8k chunks.
>
> I could be wrong, strong possibility that I am. I only got 408mb when
> doing a /dev/zero test. I've managed to saturate though. Using other
> software that I wrote.
> On 6/25/06, Nikolas Britton <nikolas.britton@gmail.com> wrote:
> > What's up with my computer, it's only getting 30MB/s?
> >
> > hostB: nc -4kl port > /dev/null
> > hostA: nc host port < /dev/zero
> >
408MByte/s or 408Mbit/s and what measuring stick are you using? I'm
trying to rule in/out problems with the disks, I'm only getting
~25MB/s on a 6 disk RAID0 over the network... would it be better to
setup an memory backed disk, md(4) , to read from?



-- 
BSD Podcasts @:
http://bsdtalk.blogspot.com/
http://freebsdforall.blogspot.com/