So Gary, if I understand correctly can I make "ordinary sieve" on sequence 9519*2^n-1 and then remove all odd candidates from sieve?
If that is true, then GPU can help us alot, since GPU sieve is way faster then CPU sieve.

[QUOTE=pepi37;477708]So Gary, if I understand correctly can I make "ordinary sieve" on sequence 9519*2^n-1 and then remove all odd candidates from sieve?
If that is true, then GPU can help us alot, since GPU sieve is way faster then CPU sieve.[/QUOTE]

Yes. Just sieve to some low limit and then remove all of the odd n's.

Alternatively you can sieve 9519*4^n-1 at half the n-range, i.e. 2^23-2^25 (n=8388608 to 33554432). No n-removal is then needed. But I don't know if a GPU can do anything except base 2.

Coordinate with Peter before doing anything. He has probably already removed odd n's or is sieving base 4. He is the starting point for all of this.

[QUOTE=gd_barnes;477709]Yes. Just sieve to some low limit and then remove all of the odd n's.

Alternatively you can sieve 9519*4^n-1 at half the n-range, i.e. 2^23-2^25 (n=8388608 to 33554432). No n-removal is then needed. But I don't know if a GPU can do anything except base 2.

Coordinate with Peter before doing anything. He has probably already removed odd n's or is sieving base 4. He is the starting point for all of this.[/QUOTE]

Very interesting :)
I will not start anything , dont worry :)
But idea with GPU still stay...

I just sent the file to Mr. Bitcoin, odd n values were removed and P=160e12. As the optimal sieve depth increases with increasing n values I wasn't sure about 25e15. But now I realize it's so much more than what I had for the previous file which went up to n=2^24 I cannot imagine 25e15 to not be enough. Well, we will see.

Any help about using GPUs will be greatly appreciated as nobody wants to waste time.

Sierpinski odd n's testing
 

Hi,

k=9267 is now tested up to n = 9,003,487 (2,710,324 decimal digits) no prime, continuing...

Best Regards,
Jean

Sierpinski odd n's testing
 

Hi,

k=9267 is now tested up to n = 9,200,215 (2,769,545 decimal digits) no prime, continuing...

Best Regards,
Jean

Sierpinski odd n's testing
 

Hi,

k=9267 is now tested up to n = 9,400,135 (2,829,727 decimal digits) no prime, continuing...

Best Regards,
Jean

Sierpinski odd n's testing
 

Hi,

k=9267 is now tested up to n = 9,602,743 (2,890,718 decimal digits) no prime, continuing...

Best Regards,
Jean

Sierpinski odd n's testing
 

Hi,

k=9267 is now tested up to n = 10,003,207 (3,011,270 decimal digits) no prime, continuing...

Best Regards,
Jean

Sierpinski odd n's testing
 

Hi,

k=9267 is now tested up to n = 10,136,695 (3,051,454 decimal digits) no prime, continuing...

Best Regards,
Jean

Sierpinski odd n's testing
 

Hi,

k=9267 is now tested up to n = 10,301,167 (3,100,965 decimal digits) no prime, continuing...

Best Regards,
Jean