I just "manually" submitted a few WU in (cuz SETI was down before), I guess something got screwed up (either at my end or SETI's end) and my WU average went from 7 hrs 27 minutes to 9 hrs and 10 minutes! Damn, I just moved into first place for fastest average and now this happens. Well, Winston Smith, you have the fastest average again ... for now. But watch out, my average is gonna be coming down fast, I have three system running SETI, and the average for them are 5 hrs, 6.5 hrs and 7.5 hrs ... I am coming for you.
February 14th, 2000, 11:51 AM
Finberg
I was wondering about your average increasing NiteOwl. Now I'm in second and dropping with my avg WU of around 6 hours. Now if I just hadn't started with a 23 hour WU and a 17 hour WU from my old celery...
February 14th, 2000, 12:02 PM
NiteOwl
your WU lookes good, what kinda system you running? Athlon?
I suggest, that is if there is a way, creating an alter ego within the group that you would use to just post your tweeked times to. This way, while you are "F"ing with the system trying to find that perfect system you don't post the obtuse times that are outside of the control limits. Use a prefix to your member name to identify the times. Something like S_USERNAME; this would be a true speed comparison. Any thoughts on the matter???
February 14th, 2000, 07:28 PM
Machupo
how do you guys get your times so low? i'm using the Win NT Commandline version on Win2k (not as a service) on my Athlon 500@800 (1/2 cache)
are there any articles out there on SETI optimization for either the NT commandline or Win2k? i couldn't find the Arstechnica one...
thanks!!!
------------------
If you get a cryptic email, this is why: 0xF2EFBF0B
February 14th, 2000, 07:49 PM
Finberg
I'm using GUI version 2.0 I tried the commandline version one time but it added 30 minutes to my time so I decided to stick with the GUI version.
February 15th, 2000, 03:17 AM
NiteOwl
I run both the GUI 2.0 or the commandline v.2 on w98se.
February 15th, 2000, 01:33 PM
millenko
just remember that each WU requires differing amounts of processing, this could result in your 1 test with the CLI version taking longer than you expect. In my experience, I have found that the CLI version is significantly faster than the GUI version, on my machine around .5-1 hour less...also depending on the amount of time that you have the graph displayed
February 15th, 2000, 01:49 PM
diablovision
I don't suggest it, and haven't done it, but you can actually edit the state.txt file and change the entry for cpu=whatever to the number of seconds of your choice.
If you stop seti at about 98% and edit the cpu= to something like an hour or 2 seconds or whatever, and then restart it, you can improve your average slowly.
Some dirty mofo could also do this from the very beginning to make it look like he's got some serious hardware...
I think they fixed this with version 2.0, so you can't cheat like that anymore.
