Re: 3 issues
On Wed, 6 Nov 2002, paul shields wrote:
> When starting the Win32 machine with the -w 64K option I get a warning and
> the window size does not change. This worked fine under 1.2
This is a known issue with all versions post 1.2, will befixed in next
release.
> Whenever I start a UDP test with the Windows machine as the server and the
> Solaris machine as a client, I get the following error
> D:\installers\iperf-1.6.3-win32>iperf -s -p 10000 -u -i 1
> Why the strange packet loss number at the beginning of the test?
Is this only with this configuration? First thoughts is an uninitialized
variable that would lead to MAX_INT on Solaris and thus not be 0 which may
have been expected. Is this only in 1.6.3 or all?
> Finally, why can't iperf be consistent reporting intervals. For example the
> following run generates a warning. This happens most frequently on slower
> connections where the performance may drop close to zero for a given report
> interval.
Well periodic reporting is very cumbersome and therefore we require an
interval higher than 500ms or .5s between reports. The inconsistency is
that reporting is not done in some off thread that has timers fire every
interval and then spits to the screen. I believe I know what the problem
is here with where the next interval gets choosen. I will fix it in the
next release if I am right. I would not recommend a interval smaller than
1 sec, if you have some special need to have ultra fine tuned reporting
get back to me and I may look into ways for you to patch iperf to do that
better at the expense of other things.
> Can the reporting accuracy be increased to at least two decimal places for
> reports in MBits
We will think about this, but I think that is a good idea. Probably be in
that infamous next release I keep talking about.
Kevin
--
Mitch Kutzko | mitch --at-- dast.nlanr.net | mitch --at-- ncsa.uiuc.edu | 217-333-1199
http://hobbes.ncsa.uiuc.edu/