Message boards :
Number crunching :
AK V8 ported release ap. issues, install, questions etc.
Message board moderation
Previous · 1 . . . 4 · 5 · 6 · 7 · 8 · 9 · 10 . . . 19 · Next
Author | Message |
---|---|
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
... I think it's the known issue with the stock app.. and the opt. app come from the stock app, so the opt. app have the bug also..?! Possible, can't be sure until the third result is in, as I thought the stock bug was a minor Vista/Stockv6 thing, but don't know much about it. I think it is more likely something weird with that new host running stock. Be interesting to see what a third host returns, as those results aren't even a close match AFAICT. Jason "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
... I think it's the known issue with the stock app.. and the opt. app come from the stock app, so the opt. app have the bug also..?! It's Kim Vater's Q6600 with Average turnaround time 7.95 days.. so we must wait around one week for the 3 result.. ;-D What you think about this result, stock or opt.? resultid=843103303 It's the Client error - Compute error from my first post.. EDIT: I see a lot of wingman with stock app and overflow, but the third and my result is well.. |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
/// I think that's a simpler problem. Looking at the debug output I can see that the host is trying to run AK_v8_Win_SSE3 ... Not a good idea on that cpu if the features in this article about Athlon 64 FX-53 are correct ... as it says it only has SSE2 Jason "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
I was little searching and found two more: stock app - overflow my PC with V8 - well result wuid=264869610 wuid=264869930 the other I posted before: wuid=264426328 |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
I was little searching and found two more: Well maybe there is something with stock, dunno, could be those machines too. I've heard it mentioned before that the optimised are more robust in some specific cases but haven't looked that much into the reasons with 5.27/5.28 Jason "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
I was little searching and found two more: To save you looking - the stock overflow stuff is Overflow (-9) with 5.27 under Vista OS. It's specifically a problem with counting/reporting 31 pulses, so all three of these qualify. Perhaps someone ought to tell Matt, because he's put a lot of effort into tracking down data noise (e.g. the radar problem), and monitors the overflow rate here. 6.4% as I type, whidh seems high - I wonder how many of them come from this bug? |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
To save you looking - the stock overflow stuff is Overflow (-9) with 5.27 under Vista OS. It's specifically a problem with counting/reporting 31 pulses, so all three of these qualify. oh so that's what Joe was spewing about ... Well that's not his fault, though he has been kicking himself every now and then. Ignore that .. it IS MS's fault ... "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
I found now two results (from different WUs) with overflow with V8.. and the wingmen have Client error - Compute error with stock app. So we must wait for the third result/s ;-) ..I trust the V8 app and believe the third result/s will be overflow also.. :-) Somebody know why the stock app have the Client error - Compute error? wuid=264625003 wuid=264951342 Every day some new experiences.. ;-D |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
I found now two results (from different WUs) with overflow with V8.. and the wingmen have Client error - Compute error with stock app. One of those looks a bit like the computer in question might have some sort of antivirus or similar interfering causing files to appear missing or inaccessible when the app tries to write. The other one looks like it might be the old graphics crash... haven't seen those for a while ... oh wait, I don't use graphics ;D "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
I couldn't imagine, that overheat of a CPU could make overflows.. Maybe to aggressive OC? Maybe it have something to do with the V8 64bit app? Or the WinXP 64bit OS? The end-result of OC is independent from the OS (32 - 64bit), or? I thought if the OC isn't well, it could make maybe Client error - Compute errors.. If my OC was to high, I had reboots or frozen BOINC/Windows (32bit).. but all the time well results.. Some examples from John Galt 007's Q6600-B3: wuid=259244074 wuid=259243971 wuid=257876125 wuid=259243523 wuid=259243394 wuid=259242958 |
archae86 Send message Joined: 31 Aug 99 Posts: 909 Credit: 1,582,816 RAC: 0 |
I couldn't imagine, that overheat of a CPU could make overflows.. Speed of overclock and temperature of CPU are two variables both of which govern whether the particular sample of the CPU you own is going to have enough speed margin to get the right answer in all cases. It is mistaken to treat either alone. Maybe it have something to do with the V8 64bit app? There is no permanent priority order of which things are most challenging for an overclocked part. All you are saying is that with a past version of the science ap your particular CPU found just running Windows in your system more challenging than running the science ap. But with another, revised, science ap, there can easily be something done in the ap which your CPU finds a bit more challenging than anything Windows makes it do in our system--in that case the science ap will fail at a lower frequency than the system. And, no, there is no rule that the symptom will always be the same, certainly not across varying samples and designs of CPU, and still less with a whole new release of the application. |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
I couldn't imagine, that overheat of a CPU could make overflows.. Uhh.. maybe it's better not to OC..?! ;-) EDIT: So overflows would be (it's) a sign for bad OC?? ..if the wingmen are well ;-) But it could be the known issue with the app also.. ;-D (I saw the overflows with WinXP also, not only with WinVista) |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
A nice example for the buggy stock app.. not only WinVista-bug.. like I said.. ;-D wuid=269740839 WinXP and WinServer2003 also.. So my one and only overflow-mistake was the known issue with the overflow-bug.. Since then no mistake.. So - OC is stable like a rock.. :-D |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
3 times AK V8 result for this WU and pending Credits?! The 4th result will be from AK V8 also.. ;-D wuid=267143201 |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
An opteron vs p4 vs Core2 , All agree there's one spike, but I guess that one must be borderline somehow. As two are running identical code I can only suggest platform variation or other random effects there. I think you'll find this kind of situation extremely rare. Jason "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
John R. @ SETI.USA Send message Joined: 3 Jul 99 Posts: 13 Credit: 82,149,277 RAC: 47 |
I'm running a Q9300 just a tad O/C'd vat 2.8 Gzh. I'm using the optimized SSE4.1 application. I just got beaten by about 200 seconds by a wingman running a Q9300 at stock speed and using the SSSE3 app. Thinking about trying the SSSE3 app myself I'm wondering what all you experts think? I've been crunching since '99, but still think of myself as a newbie when it comes to computers as compared to what I read on this board. |
DaveSun Send message Joined: 17 Jun 00 Posts: 110 Credit: 13,713,289 RAC: 2 |
I'm running a Q9300 just a tad O/C'd vat 2.8 Gzh. I have 2 Identically configured Q9300s both are on XP pro X64. One is running SSE4.1 and the other is running SSSE3X. One is faster than the other on different angle ranges but overall they seem to be pretty well matched. |
JDWhale Send message Joined: 6 Apr 99 Posts: 921 Credit: 21,935,817 RAC: 3 |
I'm running a Q9300 just a tad O/C'd vat 2.8 Gzh. You might want to keep some records of CPU times vs. AR and try both clients(SSSE3 & SSE4.1), then decide for yourself which is better/faster. With such variables coming into play as memory speed vs. CPU speed and latencies, each system is likely to have slightly different results. With the Q9300's "limited" L2 cache, this might be one chip that performs better without OCing unless the memory subsystem is finely tuned. Also you might find one client performing better at some ARs and vice-versa. Of course, you could just relax... knowing that whichever client you've chosen is much faster than the "stock" client. P.S. I went to high school with a John R. Bailey. You're not from northern California by chance? Class of '76 ? |
John R. @ SETI.USA Send message Joined: 3 Jul 99 Posts: 13 Credit: 82,149,277 RAC: 47 |
Nope, I'm not that John R. Graduated High School in 1965 in Richfield, Minn. Only time I was ever in California was 1971. Returning from 'NAM......Spent all of 4 hours there. Kinda in a hurry to get back to wife in Alabama. My brother lived in CA for a while. He taught at San Jose State for about 5 years. Now he's a networking guru for Intel. Lives up in Tualatin now. Thanks for the advice. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Or some validator bug. It would be very useful if some of SETI team at Berkeley check this WU with standart app. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.