|
*If HTML and/or UBB Code are enabled, this means you can use use HTML and/or UBB Code in your message.
Name | Post |
Cruncher |
posted 04-18-2001 16:56 ET (US)
I've noticed some strange behavior with this client compared to Seti and GammaFlux. For instance, I had one work unit process to 99% in 13 hours, then it processed the entire thing over from the beginning while staying at 99% for another 14 hours, and it never finished the unit. It got up to 555 million conformers while staying at 99% for days. What made this worse was logging off and logging back on into Windows 2K, the UD client started over from 99% and zero conformers processed each time I logged on! It didn't save any of the processed data that it wasted about 3 days working on whenever I logged off Win2K and logged back on. Finally to fix it (I assumed the program had screwed up at this point) I uninstalled the UD client and then reinstalled it, relogged back into the client with my user info, got the new data and started over again. It seems to be working OK after doing this, but I've noticed other similar strange behavior with this client like I just described. Just FYI. Seems to me this client is buggy right now. "jerrya" on UD Team 3DNow |
sandorski |
posted 04-27-2001 12:54 ET (US)
That's what they call a MFH(Molecule From Hell). Some of the molecules are far more complex, than the others. So, as you say, some will go by quickly, some take hours to days. As for re-starting at 0 conformers, it's because the UD Agent only saves completed work(molecules). Sucks, don't it? FYI, I've had a few 48+ hour ones myself. ------------------ |
Cruncher |
posted 04-27-2001 20:10 ET (US)
Ah, interesting. The worst time I've had so far has been 24 hours. This is on a 900MHz T-Bird. The worst one that screwed up (imo) is that one I described. I still think the program screwed up, because if I remember right it was stuck at 99%, 500 million conformers for three straight days. Like way over 72 hours. My opinion is the client would still be working on that packet at 99% today if I hadn't intervened (Billions of conformers? What's next, a one-year molecule?) This system isn't THAT slow. It processes SETI client WUs in about 7 hours, does 27K rays/sec in Dcypher.net's Gamma Flux client (oh yeah, "Process-Tree" now, haha). The thing I don't like about the UD client is the lack of info it gives about your stats, i.e., typical (average) processing times, etc. You get a bunch of stuff about "points generated" and "total CPU time", then you have to divide by the units you've done. And I'm assuming the "points generated" are related to hits in matching conformers. Doesn't correlate to units done or time, that's for sure. Pretty bizarre the way they came up with this compared to Dcypher (er, Process-Tree, heh) and SETI. Looks like they took the best of those two, and got weird with it. Well, weird is a way to distinguish it, I guess. ---------------- [This message has been edited by Cruncher (edited April 27, 2001).] [This message has been edited by Cruncher (edited April 27, 2001).] |
All submissions are copyrighted by their respective authors and are not for re-use in any form without their explicit written consent.
Powered by: Ultimate Bulletin Board, Version 5.37
© Madrona Park, Inc., 1998-1999.
Membaca review slot online bisa menjadi panduan yang sangat berguna. Review sering kali memberikan informasi tentang RTP slot tertinggi dan pengalaman pemain sebelumnya. Dengan memilih slot gacor yang telah direkomendasikan, peluang untuk menang dapat meningkat. Jadi, jangan lewatkan untuk memeriksa ulasan sebelum mulai bermain.