"PRP without proof (not generated)"

CuriousKit · 2023-02-10, 01:57

My last problem was a delayed upload that resolved itself, but this one has gone for 3 days now. I probably shouldn't be overly worried, but I recently completed evaluating M(114,960,997) using my Windows machine that I've been doing for the past 2 to 3 years. For some reason though, it says no proof file was generated. Unfortunately the detailed logs of that timeframe had already been cleared from the window by the time I took a look at the exponent status, and prime.log just lists the residue:

https://www.mersenne.org/report_expo...4960997&full=1

I don't think I've done anything differently compared to any of the other exponents I've recently completed, and nothing stands out as an error in the logs. Does this mean this has to go through a PRP double-check?

kriesel · 2023-02-10, 04:05

Quote:

Originally Posted by CuriousKit

For some reason though, it says no proof file was generated. ...

I don't think I've done anything differently compared to any of the other exponents I've recently completed, and nothing stands out as an error in the logs. Does this mean this has to go through a PRP double-check?

It happens sometimes. I've had generation failures related to using temporaries storage elsewhere.
Let's see whether a deeper P-1 yields anything. If it doesn't, another PRP with proof gen will be in order as DC.

CuriousKit · 2023-02-10, 11:02

I did have problems with my hard drive becoming full on that computer, but I resolved that problem months ago (it's a development computer so sometimes it gets filled with debug files and disassembly dumps). Still, I know it can cause problems sometimes. Maybe it failed to reserve enough space during initialisation and hence could never store the intermediate residues. Oh well, thanks for the answer.

kriesel · 2023-02-10, 15:37

Quote:

Originally Posted by CuriousKit

I did have problems with my hard drive becoming full

That's probably why then.
If enough temporaries are lost, it can drive the recoverable proof power from what remains to zero. I had a gpuowl run for some yet undetermined reason cease saving temporaries at 1/3 done. So no proof generation possible.

2023-02-10, 01:57	#1
CuriousKit "J. Gareth Moreton" Feb 2015 Nomadic 6B₁₆ Posts	"PRP without proof (not generated)" My last problem was a delayed upload that resolved itself, but this one has gone for 3 days now. I probably shouldn't be overly worried, but I recently completed evaluating M(114,960,997) using my Windows machine that I've been doing for the past 2 to 3 years. For some reason though, it says no proof file was generated. Unfortunately the detailed logs of that timeframe had already been cleared from the window by the time I took a look at the exponent status, and prime.log just lists the residue: https://www.mersenne.org/report_expo...4960997&full=1 I don't think I've done anything differently compared to any of the other exponents I've recently completed, and nothing stands out as an error in the logs. Does this mean this has to go through a PRP double-check?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
"PRP proof never uploaded", but proof file was generated	CuriousKit	Software	4	2022-12-20 12:42
"Data MD5 mismatch" error uploading proof file	bplenhart	PrimeNet	2	2022-12-14 01:49
What does "Error getting version number from proof header" mean?	AwesomeMachine	Information & Answers	2	2022-05-27 09:25
"PROOF" OF BEAL'S CONJECTURE & FERMAT'S LAST THEOREM	Awojobi	Miscellaneous Math	50	2019-11-02 16:50
Kladner's "Garo-proof" Dylan thread	davieddy	Lounge	2	2012-01-09 04:56

2023-02-10, 11:02	#3
CuriousKit "J. Gareth Moreton" Feb 2015 Nomadic 153₈ Posts	I did have problems with my hard drive becoming full on that computer, but I resolved that problem months ago (it's a development computer so sometimes it gets filled with debug files and disassembly dumps). Still, I know it can cause problems sometimes. Maybe it failed to reserve enough space during initialisation and hence could never store the intermediate residues. Oh well, thanks for the answer.