[Beowulf] Re: benchmarking network and performance

Jim Lux james.p.lux at jpl.nasa.gov
Thu Jan 15 10:13:13 EST 2004


----- Original Message -----
From: "Robert G. Brown" <rgb at phy.duke.edu>
To: "Jim Lux" <James.P.Lux at jpl.nasa.gov>
Cc: "Alvin Starr" <alvin at iplink.net>; <dwu at swales.com>;
<beowulf at beowulf.org>
Sent: Thursday, January 15, 2004 5:48 AM
Subject: benchmarking network and performance


> On Wed, 14 Jan 2004, Jim Lux wrote:
>
> > Thanks.. I suspect that measurements vs packet size are of some
interest.
> > (something like HINT, except for size instead of computation).
>
> netpipe.  netperf, too, but nobody loves it any more.  netpipe will let
> you transparently test raw TCP, PVM and MPI connections side by side,
> for varying packet sizes, with streams of packets long enough to get
> decent statistics on them, and autogenerate a nice little report on
> same.
>
> Most of what you learn from such a test on ethernet is that small
> packets are latency bound and the interface sends as many of them per
> second as it can (given the latency) largely independent of packet size.
> Consequently bandwidth grows linearly with packet size.  At some point
> the volume of data being transmitted STARTS to actually run into the
> wire bandwidth (if you like, the linear growth curve gets close to the
> wirespeed saturation/constant line) and the linear growth curve gently
> bends over and saturates sort-of-exponentially, typically at some 90% of
> wirespeed (for a decent interface), typically when packet sizes reach
> roughly 1K in size (maybe 2/3 of the standard MTU).


Of course, since it's a wireless channel, the "wire bandwidth" is somewhat
ill defined, at least in terms of a classical shared media ethernet type
situation. 802.11 is also half duplex, so even though the wirespeed (the
over the air bit rate) is 11 Mbps, the actual through put could be quite a
bit less, especially if it's an "ack/nak" kind of protocol being used in the
transfer. UDP could be a LOT faster than TCP, for instance, because the UDP
packets could just stream on out.

>
> Sometimes there is interesting "structure" in the curve -- "often" and
> "egregious" structure for poor interfaces, less often for good ones but
> sometimes you'll see something like dips or peaks around certain packet
> size boundaries.
>
> Oh, and lmbench has a variety of network latency and bw tests (tcp and
> udp and localhost) mixed in with the rest of its microbenchmarks.  Maybe
>
Thanks for the suggestions

Jim

_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list