mersenneforum.org SUMOUT Error Occurred.
 Register FAQ Search Today's Posts Mark Forums Read

2018-10-13, 02:29   #1
irowiki

Sep 2018

4516 Posts
SUMOUT Error Occurred.

Quote:
CPU:
click for cpu
Worker windows:
click for workers

Brand new-ish gaming laptop, thought I'd throw prime on it more of a hardware test than anything else, out of all 50~ of my machines this is the first time running into this.

This is also my first time doing P-1, everything else has been LL and LL DC.

Cooling isn't an issue, under full load it never passes 70C / 160F.

I had given Prime 10 gigs of ram, thought that was the issue so reduced it to 3, and it didn't help.

I am running a Torture test overnight to see what happens, but thought I would post first.

2018-10-13, 13:53   #2
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

160F16 Posts

Quote:
 Originally Posted by irowiki CPU: click for cpu Worker windows: click for workers Brand new-ish gaming laptop, thought I'd throw prime on it more of a hardware test than anything else, out of all 50~ of my machines this is the first time running into this. This is also my first time doing P-1, everything else has been LL and LL DC. Cooling isn't an issue, under full load it never passes 70C / 160F. I had given Prime 10 gigs of ram, thought that was the issue so reduced it to 3, and it didn't help. I am running a Torture test overnight to see what happens, but thought I would post first.
You could also consider a PRP with Gerbicz check or a PRP DC with Gerbicz check.

2018-10-14, 03:27   #3
irowiki

Sep 2018

3·23 Posts

Quote:
 Originally Posted by kriesel You could also consider a PRP with Gerbicz check or a PRP DC with Gerbicz check.

Ah, so set my work preference to that, unreserve what I have and see how it does with the Gerbicz?

I've had memtest, dell diagnostics, and a Torture test running since yesterday, and it has passed all of that.

Thank you!

 2018-10-15, 06:59 #4 irowiki   Sep 2018 6910 Posts Do I need to wait the Whole 30 days or does the PRP DC have the equivalent of a Jacobi error check?
2018-10-15, 07:19   #5
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

160F16 Posts

Quote:
 Originally Posted by irowiki Do I need to wait the Whole 30 days or does the PRP DC have the equivalent of a Jacobi error check?
Jacobi check has a 50% chance of detecting an error if one occurs.

Gerbicz check error detection is almost 100%. So, not equivalent, but in a good way; probability of lack of error detection is far better!
Error detections would show up in the results file.
Following is an example of what that might look like, from a slow system I had running for a while that became unreliable during a lengthy run.

Code:
[Sun Jul 15 12:25:15 2018]
ERROR: Comparing Gerbicz checksum values failed.  Rolling back to iteration 5249876.
Continuing from last save file.
Trying backup intermediate file: p23R1809.bu
It's a check periodically along the way.

Last fiddled with by kriesel on 2018-10-15 at 07:22

2018-10-16, 02:25   #6
irowiki

Sep 2018

3×23 Posts

Quote:
 [Oct 15 20:21] Iteration: 5460000 / 77948267 [7.00%], ms/iter: 28.953, ETA: 24d 06:58

No errors, it's chugging along. Thank you for your help!

2018-10-22, 06:06   #7
irowiki

Sep 2018

6910 Posts

I'm back! Doing a PRP DC now. Got all the way to 27.22% without any errors (no rebooting) and got a round off error:

Quote:
 [Oct 21 02:04] Iteration: 21219042/77948267, Possible error: round off (0.4375) > 0.42188 [Oct 21 02:04] Iteration: 21220000 / 77948267 [27.22%], ms/iter: 28.167, ETA: 18d 11:51 [Oct 21 02:04] Possible hardware errors have occurred during the test! 1 ROUNDOFF > 0.4. [Oct 21 02:04] Confidence in final result is excellent.
This laptop ran memtest for a few days straight, and then the prime95 stresstest for a few days straight.

Do y'all think this is more memory or processor related?

 2018-10-22, 09:00 #8 henryzz Just call me Henry     "David" Sep 2007 Cambridge (GMT/BST) 2·3·5·197 Posts Is this near a FFT boundary? If so this happens occasionally. What is the average error? Last fiddled with by henryzz on 2018-10-22 at 09:00
2018-10-22, 18:10   #9
irowiki

Sep 2018

3·23 Posts

Quote:
 Originally Posted by henryzz Is this near a FFT boundary? If so this happens occasionally. What is the average error?
That error I posted was the only one. So I shouldn't worry quite yet?

2018-10-23, 13:57   #10
henryzz
Just call me Henry

"David"
Sep 2007
Cambridge (GMT/BST)

2×3×5×197 Posts

Quote:
 Originally Posted by irowiki That error I posted was the only one. So I shouldn't worry quite yet?
Only if it is near an FFT boundary. I am not sure how best to determine that.

2018-10-24, 02:40   #11
axn

Jun 2003

53×41 Posts

Quote:
 Originally Posted by henryzz Only if it is near an FFT boundary. I am not sure how best to determine that.
If it is doing a PRP test with Gerbicz Error Check, I wouldn't worry about it. Also note the "Confidence in final result is excellent."

 Similar Threads Thread Thread Starter Forum Replies Last Post esqrkim Hardware 7 2010-06-13 17:18 mush777 Software 13 2007-04-02 17:26 garo Hardware 6 2004-12-12 15:06 Lieven Hardware 9 2004-01-03 00:38 weatherboy Software 12 2003-03-11 06:20

All times are UTC. The time now is 21:41.

Sat Sep 25 21:41:15 UTC 2021 up 64 days, 16:10, 0 users, load averages: 1.16, 1.27, 1.38