bug-wget
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-wget] mistakes in performance-view


From: Tim Ruehsen
Subject: Re: [Bug-wget] mistakes in performance-view
Date: Mon, 09 May 2016 11:23:02 +0200
User-agent: KMail/4.14.10 (Linux/4.5.0-2-amd64; KDE/4.14.14; x86_64; ; )

Hi Andreas,

please always answer to the mailing list (optionally put me on CC, but not 
really needed).

Wget outputs one line every 50k of data. On fast lines this makes many 
"measurements" / calculations per second. A slight jitter in packet delivery 
(e.g. CPU or IO busy for 0.1s) leads to totally unreliable values. It is just 
a 'best guess' from wget - from what it sees right at that moment.

What you want or need might be
--progress=dot:mega
(see man wget)

That should average the calculations good enough.

Tim

On Sunday 08 May 2016 21:43:10 Andreas Matthus wrote:
> Hallo Tim,
> 
> yes, the summery is O.K., but the values in succession arn't.
> 
> I make a wget-connection in our campus with a filesize a litle bit over
> 1 GB. I want to see the performance over all parts of the transfer - not
> only the summery. To preclude teh disk-speed the target is /dev/null.
> The pakages go over a lot of switches an cables. Some of them only use
> 100 Mb/s, but on the endpoints of the test-infrastructure are 1000
> Mb/s-networkcards and switches. In this way the maximal flow-rate is
> capped by 100 Mb/s - with possibility of a little variance by caching in
> the switches. But "little" is not so big I see in the messure. Is this
> understandable?
> 
> With regards
> Andreas
> 
> Am 07.05.2016 um 19:45 schrieb Tim Rühsen:
> > Hi Andreas,
> > 
> > just taking the values from the your attached log:
> > 
> > $ awk "BEGIN {print (1174405120 / 1024 / 1024 / 103)}"
> > 10.8738
> > 
> > Wget said:
> > 2016-05-06 13:25:10 (10,8 MB/s) - »»/dev/null«« gespeichert
> > [1174405120/1174405120]
> > 
> > That seems to be perfectly fine.
> > 
> > 
> > Is it a 100 MByte/s or MBit/s line - this seems not clear to me because
> > you
> > explicitely tell us about 1000baseT-FD. On the other hand you complain
> > about seeing speeds >12MB(yte)/s...
> > 
> > What *exactly* do you 'measure' ? Or do you just read the value from
> > '2016-05-06 13:25:10 (10,8 MB/s) - »»/dev/null«« gespeichert
> > [1174405120/1174405120]' ?
> > 
> > Regards, Tim
> > 
> > 
> > Dipl.-Phys. Andreas Matthus
> > Netzwerkadministrator
> > 
> > Technische Universität Dresden
> > Fakultät Architektur
> > 01062 Dresden
> > Tel.: +49 (351) 463-33909
> > Fax: +49 (351) 463-36120
> > E-Mail: address@hidden

Attachment: signature.asc
Description: This is a digitally signed message part.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]