Overclocked Athlon 3000+@2500 with no problems and errors till now.
The important bit of the out messages is:
Quote:
Unhandled Exception Detected...
- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x00458F49 write attempt to address 0x010ECA88
The only times I've encountered that error has been when another process has been using the BOINC directory when it was crunching. Prime cases are Antivirus or Defrag programs.
- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x00458F49 write attempt to address 0x010E30A8
Could it be that your problem is simply that the new Einstein science ap simply has an infrequent event that requires a bit more speed than your CPU can provide at its current overclock?
As background: during recent overclock experiments on my C2D E6600, I found that voltage/frequency combinations which would not only boot Windows, but would run (the previous) Einstein stably for days would give a SETI error within hours. I had to raise the voltage at least four increments higher than required to run a hour stably before this symptom declined to one in a couple of days.
Are you willing to try either raising voltage a few nothces or reducing frequency a few notches to see whether you can modulate the rate of this error? I you can show on your rig that this is in fact a speed limitation, it would help clarify the prevailing assumption that all SR2 errors are coding bugs.
Here is an example of one of my SETI failures that I am sure was a speed problem of mine:
***********
Unhandled Exception Detected...
- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x3FB58BB9 read attempt to address 0x3FB58BB9
**********
for more detail, you can view the actual wu record at:
Yeah, that usually is the reuslt of the checkpoint file getting corrupted when the machine goes down unexpectedly. I've had it happen before on S4 and S5Rx but it was rare. During this debug phase on S5R2 we'll be more susceptible to this, since the app is accessing and writing to the state files much more frequently to log the debug data.
Just information about error from the beginning of this thread.
Reinstalling OS and Boinc did not help, but after lowering FSB from 278 to 270 (in total for 72MHz) solve the problem.
What "seems to be" a stable overclock for last few years (24/7) on different projects start to cause problems on S5R2.
So, new question: Is the problem in overclock, in S5R2, or maybe both?
So, new question: Is the problem in overclock, in S5R2, or maybe both?
IMO S5R2.
I have a Athlon XP 1800+ and it's no overclocked. The 4.13 client computing, and x percent stopped, the boinc debug started, but not finished. I updated the BoincMGR, that not solved the problem. The official 4.17 client have this problem too. I suspended this project temporaly.
First client error on S5R2
)
The important bit of the out messages is:
The only times I've encountered that error has been when another process has been using the BOINC directory when it was crunching. Prime cases are Antivirus or Defrag programs.
Could this be the problem in your case?
Check all logs and events and
)
Check all logs and events and no activity in time of error.
That host is in use only for crunching.
Same or similar error
)
Same or similar error again
wuid=33357902
Unhandled Exception Detected...
- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x00458F49 write attempt to address 0x010E30A8
23/04/2007 18:40:27|Einstein@Home|Starting task h1_0158.00_S5R2__15_S5R2c_0 using einstein_S5R2 version 413
23/04/2007 22:53:52|Einstein@Home|Deferring communication for 1 min 0 sec
23/04/2007 22:53:52|Einstein@Home|Reason: Unrecoverable error for result h1_0158.00_S5R2__15_S5R2c_0 ( - exit code -1073741819 (0xc0000005))
23/04/2007 22:53:52|Einstein@Home|Computation for task h1_0158.00_S5R2__15_S5R2c_0 finished
23/04/2007 22:53:52|Einstein@Home|Output file h1_0158.00_S5R2__15_S5R2c_0_0 for task h1_0158.00_S5R2__15_S5R2c_0 absent
23/04/2007 22:53:52|Einstein@Home|Starting h1_0158.00_S5R2__14_S5R2c_0
RE: Unhandled Exception
)
Could it be that your problem is simply that the new Einstein science ap simply has an infrequent event that requires a bit more speed than your CPU can provide at its current overclock?
As background: during recent overclock experiments on my C2D E6600, I found that voltage/frequency combinations which would not only boot Windows, but would run (the previous) Einstein stably for days would give a SETI error within hours. I had to raise the voltage at least four increments higher than required to run a hour stably before this symptom declined to one in a couple of days.
Are you willing to try either raising voltage a few nothces or reducing frequency a few notches to see whether you can modulate the rate of this error? I you can show on your rig that this is in fact a speed limitation, it would help clarify the prevailing assumption that all SR2 errors are coding bugs.
Here is an example of one of my SETI failures that I am sure was a speed problem of mine:
***********
Unhandled Exception Detected...
- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x3FB58BB9 read attempt to address 0x3FB58BB9
**********
for more detail, you can view the actual wu record at:
result 521942914
Just had my first R2 error -
)
Just had my first R2 error - h1_0191.15_S5R2__3_S5R2c - let's see how the next two crunchers fare with this wu.
Ugh, I see one of your
)
Ugh, I see one of your partners had a disk space error.
Probably means my PIII will fault, since the disk BOINC is on is a bit full at the moment.
Looks like an emergency service trip will required to avoid that.
Alinator
After a power failure this wu
)
After a power failure this wu was out of the game :(
http://einsteinathome.org/task/83442924
Anders n
Yeah, that usually is the
)
Yeah, that usually is the reuslt of the checkpoint file getting corrupted when the machine goes down unexpectedly. I've had it happen before on S4 and S5Rx but it was rare. During this debug phase on S5R2 we'll be more susceptible to this, since the app is accessing and writing to the state files much more frequently to log the debug data.
Alinator
Just information about error
)
Just information about error from the beginning of this thread.
Reinstalling OS and Boinc did not help, but after lowering FSB from 278 to 270 (in total for 72MHz) solve the problem.
What "seems to be" a stable overclock for last few years (24/7) on different projects start to cause problems on S5R2.
So, new question: Is the problem in overclock, in S5R2, or maybe both?
RE: So, new question: Is
)
IMO S5R2.
I have a Athlon XP 1800+ and it's no overclocked. The 4.13 client computing, and x percent stopped, the boinc debug started, but not finished. I updated the BoincMGR, that not solved the problem. The official 4.17 client have this problem too. I suspended this project temporaly.
esattila
( Sorry for my poor english. )