Factor found that should have been found by P-1 - Page 2

Madpoo · 2015-02-20, 22:26

Quote:

Originally Posted by Prime95

The v4 server did not keep the detailed log that v5 server does. So, no, there is no way to tell which user found a factor, returned a no-factor result, did the P-1, etc. The LL results table did record the userid and user-provided computer name, but did not record the date of the result (so no help in rediscovering milestone dates).

The reason I brought up the v4 server data, is that it could be used to find P-1 bounds of exponents that have later had a factor found. The v5 server deletes that information from the "exponents with no known factor" table, but is often still available in the detailed results log.

I peeked in the v4 database and unless I'm looking in the wrong spot, I only saw some of those types of details starting with exponent 15,541,007.

I didn't dig very deep in there though.

The "option 2" to store that data in a new table seems workable, with a view that could easily show it along with whatever other info if needed. :)

tha · 2015-03-09, 11:31

-

tha · 2015-03-11, 11:12

I found one more in the same region.

tha · 2015-03-13, 10:52

And one more in the 12M region. I am still trying to figure out if there is a way to identify the P-1 assignments that have been done by an apparent faulty machine.

tha · 2015-03-20, 23:18

Again one more in the 12M region.

manfred4 · 2015-03-21, 00:07

I had some of those cases as well, for example this one.

James Heinrich · 2015-03-21, 00:42

I just discovered that my list of missing exponents... was missing some exponents

The list has now grown from 251 to 1160 exponents, including the ones mentioned above. Future discoveries should now also be properly included.

tha · 2015-03-21, 09:39

Quote:

Originally Posted by James Heinrich

The list has now grown from 251 to 1160 exponents, including the ones mentioned above. Future discoveries should now also be properly included.

Great, now we could try to figure out if we can deduce how many and particulary which machines are the culprits, and which other assignments they did.

manfred4 · 2015-03-21, 10:16

I guess you have added a little too many exponents to that list, see here and here, I think those have been found by P-1, so why are they on that list?

James Heinrich · 2015-03-21, 14:21

Unfortunately old PrimeNet records were poorly structured, a lot of data with regards to P-1 effort was not recorded. P-1 bounds, for example. The exponent record was updated with the new bounds if P-1 was done, but the bounds used were not recorded in the history log. I have attempted to revise my PrimeNet history parsing to better guess if P-1 was actually used to find the factor and relate it to the bounds recorded for the exponent. It may not be 100% accurate but is the best I can guess. I then went through the affected exponents to purge records where it appears the factor was indeed found by P-1.

So last night's 1160 exponents is now down to 295.

VictordeHolland · 2015-03-21, 14:36

Considering the amount of LL/DC tests with bad residues, I find 295 P-1 factor misses not that much. Missing a P-1 factor is in my opinion less harmful to the project than a wrong LL residue, the computer failed to save LL/DC work. Still, If we could identify 'error phrone' machines, then that is always good.

2015-03-09, 11:31	#13
tha Dec 2002 2⁵×3³ Posts	- Last fiddled with by tha on 2015-03-09 at 11:33

2015-03-21, 00:42	#18
James Heinrich "James Heinrich" May 2004 ex-Northern Ontario 5×19×43 Posts	I just discovered that my list of missing exponents... was missing some exponents The list has now grown from 251 to 1160 exponents, including the ones mentioned above. Future discoveries should now also be properly included. Last fiddled with by James Heinrich on 2015-03-21 at 00:42

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
F12 factor found?	johnadam74	FermatSearch	16	2016-11-03 12:10
Mfaktc keeps going after a factor is found	NBtarheel_33	GPU Computing	11	2012-04-07 21:12
found this factor	tha	Factoring	4	2007-06-18 19:56
After a factor is found it keeps on going	jocelynl	Software	6	2004-08-07 01:31
Odd Reporting of a Factor Found	Reboot It	Data	3	2003-12-03 14:39

2015-03-11, 11:12	#14
tha Dec 2002 2⁵×3³ Posts	I found one more in the same region.

2015-03-13, 10:52	#15
tha Dec 2002 1540₈ Posts	And one more in the 12M region. I am still trying to figure out if there is a way to identify the P-1 assignments that have been done by an apparent faulty machine.

2015-03-20, 23:18	#16
tha Dec 2002 2⁵·3³ Posts	Again one more in the 12M region.

2015-03-21, 00:07	#17
manfred4 Mar 2014 Germany 2³×3×5 Posts	I had some of those cases as well, for example this one.

2015-03-21, 10:16	#20
manfred4 Mar 2014 Germany 170₈ Posts	I guess you have added a little too many exponents to that list, see here and here, I think those have been found by P-1, so why are they on that list?

2015-03-21, 14:21	#21
James Heinrich "James Heinrich" May 2004 ex-Northern Ontario 4085₁₀ Posts	Unfortunately old PrimeNet records were poorly structured, a lot of data with regards to P-1 effort was not recorded. P-1 bounds, for example. The exponent record was updated with the new bounds if P-1 was done, but the bounds used were not recorded in the history log. I have attempted to revise my PrimeNet history parsing to better guess if P-1 was actually used to find the factor and relate it to the bounds recorded for the exponent. It may not be 100% accurate but is the best I can guess. I then went through the affected exponents to purge records where it appears the factor was indeed found by P-1. So last night's 1160 exponents is now down to 295.

2015-03-21, 14:36	#22
VictordeHolland "Victor de Hollander" Aug 2011 the Netherlands 3²·131 Posts	Considering the amount of LL/DC tests with bad residues, I find 295 P-1 factor misses not that much. Missing a P-1 factor is in my opinion less harmful to the project than a wrong LL residue, the computer failed to save LL/DC work. Still, If we could identify 'error phrone' machines, then that is always good.