Ooops, I think I screwed something up

antiroach · 2003-07-27, 22:20

A couple days ago I checked out a small range (76.0 - 76.05) I was going to do it upto 2^60, but it would have finished too quickly (i didnt realize it would go that fast) so i decided to take it up to 2^61. I changed the factoroverride= to 61 BUT and i dont have the slightest clue why i did this, but I changed all the factor=###,59 to factor=###,60. Now that I finally realized I did this, prime95 would never trial factor from 59 to 60, it would go straight from 60 to 61. Whats the best way to fix this? If i made a new worktodo.ini from the nofactor.cmp file and refactored that whole range with factoroverride set to 61 would primenet accept these "duplicate" results? Like, primenet "knows" that a certain exponent hasnt had a factor found upto 61bits for example. However the 59 to 60 bit range for that exponent has been skipped by me by accident. If I now submit a result saying that no factor has been found to 2^60 would primenet accept this? If so then I think everything would be fine. Any ideas?

cheesehead · 2003-07-27, 23:23

1) You do NOT need to submit any "no factor to 2^60" report for any Mnumber for which you previously submitted a "no factor to 2^61" report, even if that 2^61 report was in error at the time you submitted it because 2^59 to 2^60 had not yet been done.

2) What you DO need to do is to go back to all the Mnumbers where you mistakenly skipped TFing from 2^59 to 2^60, and now actually perform that TF from 2^59 to 2^60.

Once you've done that for all the Mnumbers whose 2^59-2^60 TFing had earlier been skipped, you will have two categories of results for that range: A) Mnumbers which had no factor between 2^59 and 2^60 and B) Mnumbers which have a factor between 2^59 and 2^60.

Report only the B-category results (factor exists) to PrimeNet.

Your results for the A-category (no factor found to 2^60) will already have been assumed to have been done by PrimeNet when it received your previous "no factor to 2^61" report, so PrimeNet does not need to get another report. What needs to be changed is reality, not PrimeNet. :)

What you will have accomplished will be, for Category-A results, to have changed those "2^61" reports from being false in reality to being true in reality, thus making reality conform to what PrimeNet has "thought" or "known" all along.

For category-B results, you do need to submit the new factor-found reports to PrimeNet, so that PrimeNet knows about the factors and does not erroneously continue to "know" that there is no factor to 2^61

antiroach · 2003-07-28, 01:45

yeah i realized i should do that after i made the post. but thanks for confirming :)

2003-07-27, 22:20	#1
antiroach Jun 2003 364₈ Posts	Ooops, I think I screwed something up A couple days ago I checked out a small range (76.0 - 76.05) I was going to do it upto 2^60, but it would have finished too quickly (i didnt realize it would go that fast) so i decided to take it up to 2^61. I changed the factoroverride= to 61 BUT and i dont have the slightest clue why i did this, but I changed all the factor=###,59 to factor=###,60. Now that I finally realized I did this, prime95 would never trial factor from 59 to 60, it would go straight from 60 to 61. Whats the best way to fix this? If i made a new worktodo.ini from the nofactor.cmp file and refactored that whole range with factoroverride set to 61 would primenet accept these "duplicate" results? Like, primenet "knows" that a certain exponent hasnt had a factor found upto 61bits for example. However the 59 to 60 bit range for that exponent has been skipped by me by accident. If I now submit a result saying that no factor has been found to 2^60 would primenet accept this? If so then I think everything would be fine. Any ideas?

2003-07-27, 23:23	#2
cheesehead "Richard B. Woods" Aug 2002 Wisconsin USA 2²×3×641 Posts	Re: Ooops, I think I screwed something up 1) You do NOT need to submit any "no factor to 2^60" report for any Mnumber for which you previously submitted a "no factor to 2^61" report, even if that 2^61 report was in error at the time you submitted it because 2^59 to 2^60 had not yet been done. 2) What you DO need to do is to go back to all the Mnumbers where you mistakenly skipped TFing from 2^59 to 2^60, and now actually perform that TF from 2^59 to 2^60. Once you've done that for all the Mnumbers whose 2^59-2^60 TFing had earlier been skipped, you will have two categories of results for that range: A) Mnumbers which had no factor between 2^59 and 2^60 and B) Mnumbers which have a factor between 2^59 and 2^60. Report only the B-category results (factor exists) to PrimeNet. Your results for the A-category (no factor found to 2^60) will already have been assumed to have been done by PrimeNet when it received your previous "no factor to 2^61" report, so PrimeNet does not need to get another report. What needs to be changed is reality, not PrimeNet. :) What you will have accomplished will be, for Category-A results, to have changed those "2^61" reports from being false in reality to being true in reality, thus making reality conform to what PrimeNet has "thought" or "known" all along. For category-B results, you do need to submit the new factor-found reports to PrimeNet, so that PrimeNet knows about the factors and does not erroneously continue to "know" that there is no factor to 2^61

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Ooops....SNFS poly for primes?	schickel	FactorDB	0	2011-12-16 18:07

2003-07-28, 01:45	#3
antiroach Jun 2003 2²·61 Posts	yeah i realized i should do that after i made the post. but thanks for confirming :)