Question: If a P-1 only did S1 because it found factors, and I want it to go back and do S2 (I kept the save file) what do I put in worktodo? Pminus1 or PFactor?

[QUOTE=Dubslow;289199]Question: If a P-1 only did S1 because it found factors, and I want it to go back and do S2 (I kept the save file) what do I put in worktodo? Pminus1 or PFactor?[/QUOTE]You'll need to use Pminus1:[code]Pminus1=k,b,n,c,B1,B2[,how_far_factored][,B2_start][,"factors"][/code]Notably, you'll need to include the factors that were found in stage1 in the "factors" list, otherwise Prime95 will start running off the existing savefile, find the stage1 factor(s) again and stop the assignment. So something like:[code]Pminus1=1,2,49876543,1,500000,12000000,0,"1234567890123456789"[/code]

Wouldn't c need to be -1? Otherwise, thanks. I definitely would not have gotten the known factors part. (Decimal, right?)

Is the zero for 'how far factored' or for B2 start?
Edit: Going off the wiki article, it doesn't mention factored depth, so I'll assume B2 start.


[code]Pminus1=1,2,52567117,-1,515000,15000000,0,"10005731267083223465351","6376255264142420877595961"[/code]

[QUOTE=Dubslow;289210]Wouldn't c need to be -1? Otherwise, thanks. I definitely would not have gotten the known factors part. (Decimal, right?)

Is the zero for 'how far factored' or for B2 start?
Edit: Going off the wiki article, it doesn't mention factored depth, so I'll assume B2 start.


[code]Pminus1=1,2,52567117,-1,515000,15000000,0,"10005731267083223465351","6376255264142420877595961"[/code][/QUOTE]
From whatsnew.txt:

" Pminus1=k,b,n,c,B1,B2[,how_far_factored][,B2_start][,"factors"] "

Another way would be setting Stage1GCD=0 in prime.txt.

[code]Pminus1=N/A,1,2,52567117,-1,515000,15000000,0,"10005731267083223465351","6376255264142420877595961"[/code]
[code][Worker #1 Feb 12 17:09:48] Worker starting
[Worker #1 Feb 12 17:09:48] Setting affinity to run worker on logical CPU #1
[Worker #3 Feb 12 17:09:48] Waiting 10 seconds to stagger worker starts.
[Worker #2 Feb 12 17:09:48] Waiting 5 seconds to stagger worker starts.
[Worker #1 Feb 12 17:09:48] P-1 on M52567117 with B1=515000, B2=15000000
[Worker #1 Feb 12 17:09:48] Using Core2 type-3 FFT length 2800K, Pass1=448, Pass2=6400
[Worker #1 Feb 12 17:09:50] Error parsing comma separated known factor list near: "637625526
[Worker #1 Feb 12 17:09:50] M52567117 stage 1 is 0.0417% complete.[/code]
It's not finding the save file. I just looked, and I have no clue if I have it or not. I do not have any KeepPminus1SaveFile lines in prime.txt or local.txt, so I assume it's keeping them, but there's a lot less files than there are P-1s that I've done in the last week. (Either way though, it still can't parse the factor thingy.)

[QUOTE=lorgix;289212]Another way would be setting Stage1GCD=0 in prime.txt.[/QUOTE]Ah, excellent idea, I missed that option.

[QUOTE=Dubslow;289215]It's not finding the save file. I just looked, and I have no clue if I have it or not. I do not have any KeepPminus1SaveFile lines in prime.txt or local.txt, so I assume it's keeping them, but there's a lot less files than there are P-1s that I've done in the last week.[/QUOTE]Prime95 now (since relatively recently, maybe starting with v25) defaults to keeping Pminus1 savefiles, unless told otherwise. But note, this is P-1 assignments invoked with [i]Pminus1=[/i] worktodo lines, it does not apply to the more-common [i]Pfactor=[/i] assignments. The savefiles are the same, and if you kept one manually (or can recover it via a backup or perhaps an undelete utility like [url=http://www.piriform.com/recuva]Recuva[/url]) then you can use it.

[QUOTE=Dubslow;289210]Wouldn't c need to be -1?[/QUOTE]Oops, yes, I missed the -

[QUOTE=Dubslow;289215]Either way though, it still can't parse the factor thingy.[/QUOTE]It's a single list of factors -- the comma-separation is within the quotes:[code]Pminus1=N/A,1,2,52567117,-1,515000,15000000,72,0,"10005731267083223465351,6376255264142420877595961"[/code]The how-far-factored element (72 in my above example) doesn't really affect anything except the display of factor probability. In Pfactor= assignments it will affect what bounds are chosen; in Pminus1= assignments the bounds are fixed by you, so it's just a cosmetic thing.

[QUOTE=James Heinrich;289217]
Prime95 now (since relatively recently, maybe starting with v25) defaults to keeping Pminus1 savefiles, unless told otherwise. [U]But note, this is P-1 assignments invoked with [i]Pminus1=[/i] worktodo lines, it does not apply to the more-common [i]Pfactor=[/i] assignments.[/U] The savefiles are the same, and if you kept one manually (or can recover it via a backup or perhaps an undelete utility like [url=http://www.piriform.com/recuva]Recuva[/url]) then you can use it.
[/QUOTE]

That's the part I missed. Good to know. I guess I don't have the file.

I've been pondering the last few days, if the B1 bound being used is the same for several exponents, wouldn't there be some way to check multiple exponents at the same time? I could be wrong, but it seems that the P-1 S1 calculates a lot of the same things for each exponent, how much extra effort would there be to have say 10+ exponents and have it calculate them in parallel?

[QUOTE=bcp19;289238]I've been pondering the last few days, if the B1 bound being used is the same for several exponents, wouldn't there be some way to check multiple exponents at the same time? I could be wrong, but it seems that the P-1 S1 calculates a lot of the same things for each exponent,[/QUOTE]As soon as the stage 1 exponentiation hits the first mod(Mnumber) operation, they go their separate ways.

- - -

Exercises:

What's the tenth primorial (i.e., product of first ten primes)? What is the base-2 logarithm of 3^(tenth primorial)?

How far does the stage 1 3-exponentiation have to go before it exceeds a typical Mnumber on which GIMPSters are performing P-1? To what B1 does that correspond?

[QUOTE=bcp19;289238]I've been pondering the last few days, if the B1 bound being used is the same for several exponents, wouldn't there be some way to check multiple exponents at the same time? I could be wrong, but it seems that the P-1 S1 calculates a lot of the same things for each exponent, how much extra effort would there be to have say 10+ exponents and have it calculate them in parallel?[/QUOTE]

No. To keep the intermediate calculations from getting too big, the calculation is done modulo the Mp. This means that the calculation values cannot be reused between different Mp's.

[QUOTE=cheesehead;289242]As soon as the stage 1 exponentiation hits the first mod(Mnumber) operation, they go their separate ways.

- - -

Exercises:

What's the tenth primorial (i.e., product of first ten primes)? What is the base-2 logarithm of 3^(tenth primorial)?

How far does the stage 1 3-exponentiation have to go before it exceeds a typical Mnumber on which GIMPSters are performing P-1? To what B1 does that correspond?[/QUOTE]

[I]p[/I]n#=6469693230 when n=10. log2(3^6468683230) is beyond my current ability to calculate.

Seeing as 3^2310(fifth primorial) is 1.4128e+1102 (roughly M3660) and 3^30030(sixth primorial) = 8.9388e+14327 (roughly M47590), it's fairly easy to comprehend that 3^ 8th or 9th primorial will exceed a typical Mp.

And calculating B1 is also beyond my current ability.