new prime testing program -- anyone want to test?

[jjjjL]

It was posted in another thread but it appears that phil, one of the authors that worked a few years back to improve the sieving clients has been writing a prime testing app as well.

It's written in C so it has the potential to work on lots of platforms beyond our current group but I was wondering how it performs vs SB? Are there people interested in benchmarking and letting us know?

http://www.mersenneforum.org/showthr...?t=6650&page=3


Cheers,
Louie

[Ken_g6[TA]]

Well, here's a quick benchmark I did tonight on my Athlon XP 2200+ (1.8GHz):

Phil's phrot program:
Actually testing 635481*1048576^20480+1 (20482/49152 limbs)
635481*2^409600+1 [-244748,-353119,372855,204230] is composite LLR64=F4D8A6DFADB
1882C. (e=0.10944 (0.143278~6.95512e-16@1.053) t=2326.28s)

By comparison, using the Seventeen or Bust client, in the past I've tested similar numbers in 1122 seconds:
601467*2^409600+1 is not prime. Res64: C89279AB140D4564

So it's not useful for x86 yet. When Phil gets his PPC code working, maybe some Mac afficionado (sp?) can try it.

[tnerual]

the benchmark are not good for base 2. if you try it with other base, it will test 2 to 3 times faster than LLR. so it's not usefull for SOB except that it in a near future it will allow other processor architecture