[QUOTE=Madpoo;405061]Feel free to tweak and toy with it to whatever you know works best. I'm not a factoring person so I have no idea. :smile:[/QUOTE]

OK, underway. I just left the start level at 1 -- useless I know, but only takes milliseconds. Takes between 1.5 and 4 minutes each (depending on the range / depth) to go to 61 or 62.

You should have all the results submitted within 24 hours or so (unless the cost per hour of a cg1.4xlarge goes above $0.15 -- my bid limit). A ~$3.50 well spent in my mind... :smile:

[QUOTE=chalsall;405063]You should have all the results submitted within 24 hours or so (unless the cost per hour of a cg1.4xlarge goes above $0.15 -- my bid limit). A ~$3.50 well spent in my mind... :smile:[/QUOTE]

OK, almost finished. Everything to 2^62 has been done; six factors found:[CODE]
M5215921 has a factor: 2413927249694468929 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5219077 has a factor: 4575876017916663209 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5703091 has a factor: 4164448843683868183 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5710631 has a factor: 2309255542940618039 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5710123 has a factor: 4418891578985685793 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5737429 has a factor: 2445697007785169431 [TF:1:62*:mfaktc 0.20 75bit_mul32]
[/CODE]

I don't expect to find anything in the "to 61" candidates as I assume TJAOI would have already found them, but I'm going to run them all anyway.

[QUOTE=chalsall;405087]OK, almost finished. Everything to 2^62 has been done; six factors found:[CODE]
M5215921 has a factor: 2413927249694468929 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5219077 has a factor: 4575876017916663209 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5703091 has a factor: 4164448843683868183 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5710631 has a factor: 2309255542940618039 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5710123 has a factor: 4418891578985685793 [TF:1:62*:mfaktc 0.20 75bit_mul32]
M5737429 has a factor: 2445697007785169431 [TF:1:62*:mfaktc 0.20 75bit_mul32]
[/CODE]

I don't expect to find anything in the "to 61" candidates as I assume TJAOI would have already found them, but I'm going to run them all anyway.[/QUOTE]

I guess it's semi-official then... TomFakes was faking it. :smile: (okay, not really, but a buggy machine or something).

Were the numbers where his system reported "has a factor: 1" ones where the number does have a factor in the range he ran? That would suggest he ran the tests, but his system was reporting 1 instead of the real factor.

Chris

[QUOTE=chris2be8;405098]Were the numbers where his system reported "has a factor: 1" ones where the number does have a factor in the range he ran?[/QUOTE]

Might have been P-1 tests. I have vague recollections of a P-1 bug causing this.

[QUOTE=Prime95;405100]Might have been P-1 tests. I have vague recollections of a P-1 bug causing this.[/QUOTE]

Oh yeah, true. I forgot that with the old version 4 logs there isn't a distinction between the different types of "no factor found" for P-1 or TF. So that could be.

[QUOTE=chalsall;405087]I don't expect to find anything in the "to 61" candidates as I assume TJAOI would have already found them, but I'm going to run them all anyway.[/QUOTE]

OK, everything has now been "worked".

As expected, no new factors were found in any of the "to 2^6[01]" candidates.

Would it be worth searching the old logs for other "factor found: 1" lines and re-running P-1 (or TF) against any numbers that don't already have a factor in the database? That should give you a list of numbers that all have a small factor to find.

Chris

[QUOTE=chris2be8;405153]Would it be worth searching the old logs for other "factor found: 1" lines and re-running P-1 (or TF) against any numbers that don't already have a factor in the database? That should give you a list of numbers that all have a small factor to find.

Chris[/QUOTE]
At this rate, I think that TJAOI will eventually find all these errors.

[QUOTE=alpertron;405154]At this rate, I think that TJAOI will eventually find all these errors.[/QUOTE]

Could be.

I looked at the old v4 logs and there are 113 entries where a user reported "UID* has a factor: 1" and there's no known factor listed for them.

80 are from TomFakes, 24 from a user "dswanson" and then a handful of users with 1, 2 or 3 apiece.

If this is indeed a P-1 related bug, I can do the pm1 work on that many... they're smallish exponents for the most part (a handful, 5, in the 20M-40M range, but mostly 16M and below).

The question I have to ask myself is... do the B1/B2 values stored in the database reflect the bounds that assignment did when it found the amazing factor of "1", or were they using different bounds that are now unrecorded anywhere?

I'm trying a few using the recorded b1/b2 and see what happens... if nothing's found I may push those bounds higher and see, then figure out if it's worth doing all 113 to any particular level.

[QUOTE=Madpoo;405156]The question I have to ask myself is... do the B1/B2 values stored in the database reflect the bounds that assignment did when it found the amazing factor of "1", or were they using different bounds that are now unrecorded anywhere?[/QUOTE]

There is no need to re-run the P-1. Prime95 did not report the bounds when a factor (1 or any other) was found. The bounds in the database are for legitimate P-1 runs after the failed run.