Dear All, After searching for some week to find a program to do 100 000 000#+ Mersenne Prime with MPI, my research resulted to nothing. So, now i'm ready to pay, if someone can implement MPI to mprime : http://mersenneforum.org/gimps/source2511.zip , or to Glucas : http://www.oxixares.com/glucas/.
These two program, make Lucas-Lehmer test on an exponent to find if it is a Mersenne Prime. They both use FFT (Fast Fourrier Transform) to make the calculation. The reason why i want this, is that exponent goes bigger and bigger, and no good program to use a HPC Cluster have been made to speedup the research. Today it take a good 2 weeks on a New 4 core machine to goes through the calculation, and 1 month for a P4 3.0GHz HT. The person who will accept to do it will greatly contribute to the Mersenne research, and make some money too. I've tried to find a way to implement the communication without falling in the embarrassingly parallel, but i'm not advanced to do so. Finally, a reward i offered too by the Electronic Frontier Foundation for finding new prime, and i'm ready to share it with the one that will make it, IF i ever find one... So, if someone is interested, plesase let me know, Thank you! Mikael Lavoie, mikael.lav...@gmail.com