ISPA Technical Testing Software Oddity

[#1]

Don't know if this is the correct forum for this but I'm sure someone will shift it if not.
I was invited by PN to join in the ISPA testing for the Best Unlimited Fibre award. I installed the software yesterday and so far it has run 3 tests which seem a bit strange to me.
The routine is claiming an average of 21.2Mb download speed and 10.1Mb upload over the tests.
My line invariably reports over 73.5Mb download and over 15Mb upload on the BTW test - rock solid, never seen it drop since it was installed.
So is this software limited in some way? Or has it just been unlucky in running the tests when I've been downloading something else? Or what?
Just seems a bit odd that I'm getting these results on software which is used to decide what the quality of the supplied service is. My service is brilliant but the test software results so far are rubbish.
Maybe it'll average out OK but even these early results seem as though they could skew the result unfairly.

[#2]

See also the thread I started on "other queries" on exactly this:
http://community.plus.net/forum/index.php/topic,112730.0.html
The TBB software is definitely telling porkies.  Not calculated to inspire confidence in their results when they finally come out.
Jonathan

[#3]

Quote from: oxfordjonathan

See also the thread I started on "other queries" on exactly this:

Oops! sorry - missed that one. Maybe a moderator could move my post to your thread?

[#4]

I've dropped an email to the ISPA guys asking about the speeds, they are going to ask TBB why it could be lower.

[#5]

I understand that .net 3.5 is needed for this TBB software to run correctly. Would it therefore be reasonable to assume that the TBB package results could be affected by all this M$ bloatware on less powerful machines?

[#6]

For those seeing significantly lower results than expected, can you PM me the 'testerinstance number' from within the software?
TBB think this may be the software only displaying the single thread result but want to take a look deeper into the issue.

[#7]

Hi everyone
Andrew (MrSaffron) from thinkbroadband here, just thought I'd update directly as there is so many of you using the tester.
The problem was that the multiple thread test result was not been written to speed.csv file which is used by the system tray icon program (PickProvider.exe) to give the users feedback.  There was no problem with the testing itself, the test was been stored on our servers and adding to the testing for the awards.
An update has gone out to the TesterLibrary.dll file which will fix this, and the software may pop-up a message about a reboot or restart of the service. There is no major panic to do either, you can wait till you next reboot the PC as still a good number of weeks of testing to run.
On why a single thread gives such a different result to multiple threads, it is a very long answer, but latency can come into play combined with RWIN as well as many other factors.

[#8]

Cheers for the direct reply about that Andrew!