Re: benchmark



On Fri, Jan 05, 2007 at 09:56:31AM -0800, pete wright wrote:

Tried to use iperf from ports in UDP mode with 64 byte payload,
but it calls gettimeofday() after each write and gives me about 80Kpps only
for Pentium D 2.8Ghz.

What alternative should I use? May be, a netgraph node?

I've done some benchmarking/testing of 10gig-e NIC's using a combo of
iperf/netgraph and ttcp with good results. all are available in
ports.

What pps numbers had you obtained? What CPU had you used?
I don't like iperf for gettimeofday() overhead.

Eugene
_______________________________________________
freebsd-performance@xxxxxxxxxxx mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-performance
To unsubscribe, send any mail to "freebsd-performance-unsubscribe@xxxxxxxxxxx"



Relevant Pages

  • Re: benchmark
    ... for Pentium D 2.8Ghz. ... May be, a netgraph node? ... I don't like iperf for gettimeofday() overhead. ...
    (freebsd-stable)
  • Re: Possible bug from kernel 2.6.22 and above, 2.6.24-rc4
    ... this does not take the overhead of gettimeofday into account - which overhead can easily reach 10 usecs. ... I first call refer subroutine which actually establishes the elapsed time (reference time) for non-synchronized delayusing the gettimeofday. ... The results of pthread_sync is the same as the non-patched 2.6.21 kernel. ...
    (Linux-Kernel)
  • Re: Possible bug from kernel 2.6.22 and above, 2.6.24-rc4
    ... Could you try to eliminate the gettimeofday overhead from ... reference time from the elapsed time with introduced synchronization. ... The effect of gettimeofdayshould be minimal if the time difference ...
    (Linux-Kernel)
  • Re: Possible bug from kernel 2.6.22 and above, 2.6.24-rc4
    ... this does not take the overhead of gettimeofday into account - which overhead can easily reach 10 usecs. ... I first call refer subroutine which actually establishes the elapsed time (reference time) for non-synchronized delayusing the gettimeofday. ... Then each synchronization overhead value is obtained by subtracting the reference time from the elapsed time with introduced synchronization. ...
    (Linux-Kernel)
  • Re: How is Code in do_sys_settimeofday() safe in case of SMP and Nest Kernel Path?
    ... I really hate adding overhead to gettimeofday() and we would have to take ... resulting race? ... So we may get daylight savings time wrong. ...
    (Linux-Kernel)