here you are c10ck3r

[QUOTE=firejuggler;310785]my FermatFactor=31,600e12,650e12 mean that a factor would be of the form

k*2^33+1 right?[/QUOTE]

Correct.

Luigi

Thank you!!!
[STRIKE]Turns out I only reserved ~8 minutes worth of work[/STRIKE] :D

That was just the first bit range...15 hrs till "actual" finish...oops...Also, dropped from 720M/s to 450M/s.

done with my range :
[code]
no factor for F31 from 2^82 to 2^83 (k range: 600000000000000 to 650000000000000) [mmff 0.20mmff mfaktc_barrett89_F30_61gs]
[/code]

took a bit above 5 hours
also taking
FermatFactor=46,300e12,325e12

k*2^48+1

[QUOTE=c10ck3r;310789]Thank you!!!
[STRIKE]Turns out I only reserved ~8 minutes worth of work[/STRIKE] :D

That was just the first bit range...15 hrs till "actual" finish...oops...Also, dropped from 720M/s to 450M/s.[/QUOTE]


It's a 420 days' work for a PentiumII-400, or 30-40 days for an average CPU...

Luigi

strangely, my new range (FermatFactor=46,300e12,325e12) goes twice slower. Different kernel? (runtime around 6H)

[QUOTE=firejuggler;310797]strangely, my new range (FermatFactor=46,300e12,325e12) goes twice slower. Different kernel? (runtime around 6H)[/QUOTE]


Probably. F46 requires 46 squarings mod f, F31 only 31 squarings. This is 50% of the extra time. For the same k-value, an F46 factor will be 15 bits bigger. This probably bumped you up to the next kernel requiring wider (slower) squarings.

Now that testing each factor is slower, it will pay to sieve a little deeper. That's why we need to generate a table of best GPUSieveSize settings for each Fermat/MM number. Long-term I suppose auto-tuning needs to be added to the source code.

[QUOTE=firejuggler;310797]strangely, my new range (FermatFactor=46,300e12,325e12) goes twice slower. Different kernel? (runtime around 6H)[/QUOTE]

hi,
FermatFactor=46,300e12,325e12
runs fine on my gtx580
raw rate is about 1050M/s - ETA is about 2h45min

Implausible thought of the day: by definition, any factor of MM43112609 would be a world record prime :P

All the .dll files are located here: [url]http://sourceforge.net/projects/cudalucas/files/CUDA%20Libs/[/url]

If anyone needs anything not listed there, just let me know.

Should we start a SourceForge Page for CUDA-mmff?

AMD/ATI users miss out on a lot of things :devil:
Maybe I might get a Nvidia card :lol: