Pentium 90 // Pentium ][ 400 years

[ValerieVonck]

Dear,

Can anyone enlighten me on how to calcuate the amount of P90/ P2-400 years that I have devoted to calculate e.g.

2293*2^2404631-1

or numbers of the form k*2^n +/-1

a) it is for me personnaly to learn how to do this.
b) for the octoproth project hosted here.

I did not find anything on the internet.
I would appreciate it!

Thank you.
Cedric

[cheesehead]

What's the FFT size?

[ValerieVonck]

The fft size of these exponents?
I searched these forums and I came with these lknks:

http://www.mersenneforum.org/showthread.php?t=6471
http://www.mersenneforum.org/showthread.php?t=1598
http://www.mersenneforum.org/showthread.php?t=120

I tried to figure out but I did not anywhere the fft size
I looked at the code of Prime95, even there was no mention of "defined" lengths.

Thnx

[Thomas11]

Perhaps you should have a look into the following post:
http://www.mersenneforum.org/showpos...2&postcount=35

There is a small tool "fft_len" which predicts the FFT length for given k and nmin/nmax range. The FFT length for your specific k/n combination would be 196608 (192k).

In order to predict the cpu times per LLR test you'll need to edit the "times.txt" file and change the times/iteration (in ms) per fftlen to values specific for your machine.

Hope this helps a bit...

-- Thomas

[ValerieVonck]

Originally Posted by Thomas11

Perhaps you should have a look into the following post:
http://www.mersenneforum.org/showpos...2&postcount=35

There is a small tool "fft_len" which predicts the FFT length for given k and nmin/nmax range. The FFT length for your specific k/n combination would be 196608 (192k).

In order to predict the cpu times per LLR test you'll need to edit the "times.txt" file and change the times/iteration (in ms) per fftlen to values specific for your machine.

Hope this helps a bit...

-- Thomas

Thomas,

I will certainely look into this.
Thank you very much.

Regards
Cedric