Testing new Ranges for Sierpinski/Riesel

[KEP]

Originally Posted by pepi37

Due latest news, I so very small check
Srbsieve from some reason say that 10*708^27+1 is prime

But independently llr64, pfgw says it is composite

10*708^27+1 is not prime. RES64: A46C206476D17E99. OLD64: 02C4612D64747BC7 Time : 58.039 ms.

D:\PGFW>pfgw64 -d -q"10*708^27+1"
PFGW Version 3.7.10.64BIT.20150809.Win_Dev [GWNUM 28.6]

Switching to Exponentiating using GMP
10*708^27+1 is composite: RES64: [A46C206476D17E99] (0.0001s+0.0002s)

How far did you sieve with NewPGen? I'm asking in terms of n, what was your smallest and what was your highest n sieved using NewPGen for S708? Did you sieve to at least n=27 (or higher) using NewPGen? Did you at all n's let NewPGen sieve up to squareroot of the highest k/n pair?

I've had no problems with srbsieve finding wrong primes and if the users fail to understand how important it is to not sieve to a higher n than NewPGen can sieve within reasonable time to squareroot of highest k/n pair, then srbsieve will mistakenly believe that you have a prime at the inappropriately sieved n.

Ps. I'm now going to do my own test of srbsieve latest version and run S708 to n=2500 and compare to Gary's list.

[rob147147]

For reference, I added 10*708^27+1 to factordb, and its first factor is 18960572317

[pepi37]

Originally Posted by KEP

How far did you sieve with NewPGen? I'm asking in terms of n, what was your smallest and what was your highest n sieved using NewPGen for S708? Did you sieve to at least n=27 (or higher) using NewPGen? Did you at all n's let NewPGen sieve up to squareroot of the highest k/n pair?

I've had no problems with srbsieve finding wrong primes and if the users fail to understand how important it is to not sieve to a higher n than NewPGen can sieve within reasonable time to squareroot of highest k/n pair, then srbsieve will mistakenly believe that you have a prime at the inappropriately sieved n.

Ps. I'm now going to do my own test of srbsieve latest version and run S708 to n=2500 and compare to Gary's list.

As tutorial say ( I followed tutorial step by step) I do from N=1 to N=50 , every step sieved to 20000000088 ,and let Newpgen finished job.
Since I do on quad core: I setup four directory with ranges that are not overlapping 1-7090 7091-14180 14181-21270 21271-28360 and do for all 4 ranges same step in sieving with NewPgen
Is that mistake I made?

[KEP]

Originally Posted by pepi37

As tutorial say ( I followed tutorial step by step) I do from N=1 to N=50 , every step sieved to 20000000088 ,and let Newpgen finished job.
Since I do on quad core: I setup four directory with ranges that are not overlapping 1-7090 7091-14180 14181-21270 21271-28360 and do for all 4 ranges same step in sieving with NewPgen
Is that mistake I made?

Yes pepi37, that is exactly the mistake you have made. I can tell you, that I have no prime for 10*708^27+1, in fact k=10 still remains at n=660. You and others have to understand, that for any base and any n, you have to let NewPGen sieve up to squareroot of the highest k/n pair. If you go and set an insane stop at p= value to something like p=1000000000000000 then tell's NewPGen to stop at n=6 for S708, then you should have ok sievefiles, because then NewPGen will automatically stop sieving the given n value, once NewPGen has sieved to squareroot of given n and given k range. You have overdone NewPGen and undersieved by many years of sieving and that way your results got unreliable and caused a friendly fire against both Mark and srbsieve, while Mark nor srbsieve has done any mistake worth blaming them for.

I'm going to announce what I find, once I complete testing S708 to n=2500, but I do expect a match between srbsieve, version released 15th august 2015 and the files Gary uploaded.

Take care!

[firejuggler]

ok, now the newbase script has finished, and the number of remaining base is the same, 22034.

here is the file to compare the result : http://www.filedropper.com/compare-srbsievenewbase (size is about 2 MB)

[rogue]

Originally Posted by pepi37

As tutorial say ( I followed tutorial step by step) I do from N=1 to N=50 , every step sieved to 20000000088 ,and let Newpgen finished job.
Since I do on quad core: I setup four directory with ranges that are not overlapping 1-7090 7091-14180 14181-21270 21271-28360 and do for all 4 ranges same step in sieving with NewPgen
Is that mistake I made?

I don't know how you read that from this:

npgfile= is used to specify output files from newgpen's fixed n sieve. The first value is the n for which the range was sieved. The second is the name of the file that was output from newpgen. You can specify as many of these as you want, although there is a practical limit due to newpgnew limits with sieving. newpgen will sieve until p exceeds sqrt(maxk*b^n+c) and it will terminate on its own. That is great for this search because it means that all values remaining after sieving are prime which means that none of the values that passed sieving will require a primality test. You will not want to go to a very high n as sieving takes longer as n increases and the effectiveness of removing k is diminished. n=22 is about as far as one should go for Riesel base 3. Higher bases will result in going to a lower n.

For the sentence in red, it is imperative that you let newpgen finish on its own. You cannot terminate. If you terminate newpgen before it reaches that max, then you cannot use that input file for srbsieve. srbsieve assumes that you have followed that rule because it guarantees primality of those numbers. srbsieve will not verify that you followed the rule.

For the sentence in green, I know understand why you ran into that problem when you had more than 50 input files to srbsieve. Nobody could have that many input files and properly sieve with newpgen. sqrt(maxk*b^50) is well beyond the capabilities of newpgen when b > 2.

Please let me know if this is clear. Also ask any questions you might have about using srsieve properly.

[pepi37]

1. Please I apologize to all: especially to Mark, GDBarnes, and any other person that is affected because my unknown of mathematics, and skill I need to know to start new base. Yes, I am sorry, and I will never do it again.
2. Please unreserved all my bases: once, in time ( if came) I will take one base and try... ( but until that time..)
3. But last , I ask any of you, that know math far better then me to in answer give me
- appropriate setup for base S708 but from scratch.
So give me number of sieved files, give me setup in NewPgen and give me basic math: how you calculate those values for S708.
Can you do it?
Dont give me formula, give me formula and result so I can follow your work.

Then when I look at those results I will send to any of you who knows math setup for next base. I will not reserve it, I will not publish it, or I will not search it, just setup so you can verify or not , tell me is there error or not.
Can I ask you this favor?
Once more, please accept my apologize..
And last: I never terminate NewPgen: it always done to the end, with my setup was over in 40 minutes, so I wait until last n was finished and message appear NewPgen reach limit.,

[KEP]

Originally Posted by pepi37

1. Please I apologize to all: especially to Mark, GDBarnes, and any other person that is affected because my unknown of mathematics, and skill I need to know to start new base. Yes, I am sorry, and I will never do it again.
2. Please unreserved all my bases: once, in time ( if came) I will take one base and try... ( but until that time..)
3. But last , I ask any of you, that know math far better then me to in answer give me
- appropriate setup for base S708 but from scratch.
So give me number of sieved files, give me setup in NewPgen and give me basic math: how you calculate those values for S708.
Can you do it?
Dont give me formula, give me formula and result so I can follow your work.

Then when I look at those results I will send to any of you who knows math setup for next base. I will not reserve it, I will not publish it, or I will not search it, just setup so you can verify or not , tell me is there error or not.
Can I ask you this favor?
Once more, please accept my apologize..

No problem Pepi, I have helped Reb understand things before and I would also like to help you. It is most welcome, that we get more users (new as well as old) that start up bases. So in a few minutes I'll send you a PM and all I ask of you, is to think of wich base to test/practice on ... welcome on board my friend, looking forward to be working with you in PM or here

[pepi37]

Originally Posted by KEP

No problem Pepi, I have helped Reb understand things before and I would also like to help you. It is most welcome, that we get more users (new as well as old) that start up bases. So in a few minutes I'll send you a PM and all I ask of you, is to think of wich base to test/practice on ... welcome on board my friend, looking forward to be working with you in PM or here

KEP, I have no other word to say then THANKS
I was learned that world open even iron gate

[firejuggler]

well,708^6 has 18 digits, 708^7 has 20 digits. i would suggest to limit yourself to 6 or seven files, as rogue recommend in the first post that k*b^n+c be slightly above 1e20. so dépending of the k....

[rogue]

Originally Posted by firejuggler

well,708^6 has 18 digits, 708^7 has 20 digits. i would suggest to limit yourself to 6 or seven files, as rogue recommend in the first post that k*b^n+c be slightly above 1e20. so dépending of the k....

Maybe a simpler way would be to sieve for n until p > 1e9 when the sieve stops on its own. That would guarantee that newgpen finds all primes <= 60 bits. p > 1e10 works better for smaller bases.