small inconsistency in the "Bytes transferred" and "speed" values

To: <iperf-users --at-- dast.nlanr.net>
Subject: small inconsistency in the "Bytes transferred" and "speed" values
From: "Leese, MJ \(Mark\)" <m.j.leese --at-- dl.ac.uk>
Date: Sat, 14 Jul 2007 19:09:06 +0100
Content-class: urn:content-classes:message
Content-transfer-encoding: 8bit
Content-type: text/plain; charset="iso-8859-1"
Importance: normal
Reply-to: iperf-users --at-- dast.nlanr.net
Sender: owner-iperf-users --at-- dast.nlanr.net
Thread-index: AcfGQhFf0nLzH0ztTHmliURQ7C9VKw==
Thread-topic: small inconsistency in the "Bytes transferred" and "speed" values

Hi,

Is there a small inconsistency in the "Bytes transferred" and "speed" values reported by iperf, or am I missing something obvious :)? Here's an example....

    [root --at-- hepgrid7 ~]# iperf -c <hidden-dest> -p 5001 -P 1 -w 1048576 -t 5 -fb
    ------------------------------------------------------------
    Client connecting to <hidden-dest>, TCP port 5001
    TCP window size: 2097152 Byte (WARNING: requested 1048576 Byte)
    ------------------------------------------------------------
    [  3] local <hidden-src> port 56301 connected with <hidden-dest> port 5001
    [  3]  0.0- 5.0 sec  296804352 Bytes  474664156 bits/sec

If we've transferred 296804352 Bytes in 5 seconds, that's (296804352 / 5.0) * 8 = 474886963 bps, not 474664156. 

It's only a small difference (just 0.2 Mbps in this case) but I'd like to be able to explain it (seriously, people point it out in examples). It's especially annoying when some smart person points it out during a talk you're giving ;-)

I had a quick play with the code but couldn't see where this was being introduced. This particular iperf is...

  OS:     a re-badged version of RHEL4.2
  kernel: 2.6.9-42.0.8.ELsmp
  glibc:  glibc-2.3.4-2.25
  iperf:  pre-built iperf 2.0.2 executable from Dag Wieer's Repository (http://dag.wieers.com/apt/)

...although I've seen it several times elsewhere.

Thanks in advance for any help,

Mark.


Mark Leese

Senior Network Designer		tel: +44 (0)1925 603730
STFC Daresbury Laboratory	http://gridmon.dl.ac.uk/~mjl

Follow-Ups:
- Re: small inconsistency in the "Bytes transferred" and "speed" values
  - From: Metod Kozelj

Prev by Date: Re: Clarification on behaviour of Iperf's -w option
Next by Date: Re: small inconsistency in the "Bytes transferred" and "speed" values
Previous by thread: Clarification on behaviour of Iperf's -w option
Next by thread: Re: small inconsistency in the "Bytes transferred" and "speed" values

Other Mailing lists | Author Index | Date Index | Subject Index | Thread Index