Hi Robin,
The paragraph below is interesting as it something we ourselves have
noticed.
>Another thing I noticed was that our system of three m/cs on a loop
>kept crashing until we had logged a multiplier. This was during a pre-
>contest check. We also had 4 crashes in the first hour resulting in our
>lowest first hour rate of 105 Qs during 6 years of using NA for the
>IOTA contest. (yes, even before it supported the event!)
Our set up was 3 PC's (1 x Run, 1 x Mult and 1 x Cluster) and we certainly
had about 3 funnies where the logs got all out of step, we think we know
why, plus 2 Total crashes of the program itself unfortunately we didn't make
a note of what they were in the heat of the contest.
The logs out of step we think arises due to say the multi station grabbing a
serial number but taking along time to confirm, in between times the RUN
station has grabbed a few more QSO's and we suspect that the buffer has now
got full in the MULT PC and when finally the grabbed QSO is confirmed the
buffer has overrun in the MULT PC and QSO's are lost giving now the serial
numbers out of step.
Wish there was away that either the serial numbers can be sorted out
stopping the grabbing scenario or that the RSGB just forgets about them for
the IOTA contest. Why not get none-IOTA stations to pass RST + NI (not
IOTA).
Cheers,
Bob Glasgow GM4UYZ
Sec: C&PSARC
IOTA Contest - GM2T Isle of Tiree EU008
|