S40 Observation Thread
3 Apr 2006 8:46:32 UTC
Topic 191019
(moderation:
So I've just calculated two WUs with the new S40 and the times are slightly longer then average from the S39L. But I need higher statistics for some more serious conclusions.
Pentium 4 Prescott (16 kB L1 cache), 3.4 GHz HT
S40 Observation Thread
)
Hello! S40 did some "access violation" error on one of my PC.
So, don't use this version! I will check the code and the PC.
I have calculated two WU's on
)
I have calculated two WU's on my Pentium Xeon (Prestonia core - L1=8kb) and it seems S40 is about 3% slower as S39L. One Wu is valid and one is pending. Tommorow morning I'll have a more crunched WU's for statistics
Have 12 WU's completed on a
)
Have 12 WU's completed on a quad 700/100/1024 Xeon, about 20 minutes faster and no problems. Looks good so far :)
RE: Hello! S40 did some
)
My Pentium4 (Prescott) 3,2 GHz HT reported 2 WUs without Error, but not faster than S39L.
I changed back to S39L an waiting for the Trouble-Shooting Report.
Chris
*Die Signatur befindet sich aus technischen Gründen auf der Rückseite dieses Beitrages!*
Have somebody got access
)
Have somebody got access violation error with S40?
S39L -> S40
Duron: 4083 -> 3745 ~ 9%
Dothan: 4103 -> 3971 ~ 3%
RE: Have somebody got
)
Sorry, a little bit Off-Topic :-(
It seems that my box has troubles with C37. Until a few days ago, I used the standard client and the box was stable as a rock. But since I have C37 running, it suddenly freezes. No idea why?
Okay, when you look at the results, you see 4 Client Errors at the end of the list. These errors are all my fault, I did not use binary ftp transfer to install the C37 Albert, dumb me used text transfer :-(
However, starting with that day I had to reboot this box 5 times, the problem seems to happen when a WU reaches 100% and boinc schedules another WU, maybe I attach a photo (screenshot on a frozen box is hard) of Boincview.
I see no error messages, no entry in the event log, just a freeze. Even the keyboard is dead.
This might be caused by high temperature, but might also be caused by whatever strange thing at the very end of a WU?
Any Idea?
The C37 experts who run that
)
The C37 experts who run that app might not be monitoring this thread. They are over here.
me-[at]-rescam.org
RE: Have somebody got
)
S40- Better prefetch, AGI
)
S40- Better prefetch, AGI optimizations. Is that for the Asrock motherboards? (AGI - AGP)
RE: S40- Better prefetch,
)
AGI (address generation interlock) isn't depend on the type of motherboard.
It comes from unlucky instruction order and register usage.