very short WUs

[derek]

Last week I had 4 WUs complete in <5 min. I assumed that I was having a problem with my slightly overclocked dual PIII 1000 system since they all occured on the same CPU. On Friday I put together a 1.2 Athalon MP and Epoc 8K7T+ system. Out of the first 10 WUs, 2 finished in <1 min. I know the MPs are quick, but not that fast. Besides I haven't started OCing it yet. I was still getting a baseline time before I started tweaking things.

Anyway I started looking at the WU Names in SetiWatch and noticed they all ended with .102. I am wondering are these just very short WUs or is there something causing my setups to bail on them?

Thanks for the help.

Sincerely,

Derek

[Burning Phoenix]

I just posted in another thread here earlier about a similiar problem where I had in the last 24 hours 4 WU's that finished in 0 hours 0 minutes.

[TC]

That's not normal, and since it's doing it on a machine that's not overclocked at all I would suspect either a software glitch, or perhaps they sent out a batch of screwed up work units. The work unit file should end in .sah The only other explanation would be an overclocked machine that's past the point of stability, but not crashing the OS.

[JigPu]

Only things I know that can make a WU finish that fast (when everything is OK)...

1) REALLY high AR (somewhere in the thosands) Never EVER heard of such a high angle range (nowhere even CLOSE) so its probably #2

2) Lots and Lots of spikes. When SETI gets lots of spikes when processing, it figures that the whole WU will be like this, and be a waste of time. It stopps processing, and sends the result.

As for the .102... That is the end WU name. Each WU has its own name. The WUs ending in .102 might all be bad WUs with tons of spikes... Or you could have just got lucky.

JigPu

[Morpheus]

Haven't seen one of the above mentioned, but I've had some 7 to 9 second WUs.. RFI or other noise is most often to blame... I suppose a splitter malfunction on the SETI server might screw things up too...

oh well....

Crunch On!!!

[andz]

A few days ago had a 15 minutes W/U...
8/23/01 6:31:24 PM 12mr01aa.4042.28513.411056.153 4.855 26.990 0.721 0.121 0.139 3.13% 128 288.839 0 1.666 0 0.990 26 9.839

As you can see, only 3.13% of the W/U was completed. I think Jigpu hit the nail in the head.

[[Oc]acaridans]

Yesterday I had a 5sec WU's
Out of luck I was looking at the results of the previous WU when it switched....It seemed to update every second it had over 200 spikes and over 300 triplets returned, and it displayed it all in 1 line.

[lonewolf1983]

ive just got two 0 hours,0 minutes WU's as well
and now i cant upload em
oh well,glad i got setistash
woohoo two 'free' WU's for me
:)

[Basher]

I had one a little while ago
8/20/2001 2:01:17 AM 26fe01aa.11384.25361.348568.102 5.895 16.560 0.423 0.083 0.129 2.02% 36 292.471 0 1.504 0 0.971 73 13.699 Intel

[derek]

The IDs for 4 of the 6 I have gotten are:

02dc00aa.9280.9600.878390.102
12fe01aa.5687.16753.404820.102
08ja01aa.5692.7008.40912.102
16mr01ab.23627.26769.997154.102

JigPu may have a good point, the output for these 4 look like they have many more spikes that normal units. Some of the normal WUs look like they are almost 0s.

It looks like other people are getting these WU and short times, so I guess I will start bumping up the speed on the 2 machines.

Thank you for the help.

Sincerely,

Derek