So, guys, I notice the Tour De Primes is coming up soon. I don't have any new hardware, but I'd like to make the best use of the hardware I've got. So, does anybody know how to set up LLRNet with llrCUDA, say, for the PPSE range? Has anyone even done it? Can I do LLRNet work on the CPU at the same time?

Thanks!
____________

For now it is not so efficient, so any moder CPU will outperform CUDA card. And if you have CPU with AVX , in that case...that CUDA card is irrelevant.
But if you wont anyway to use it: then just make another directory for LLRCUDA and uncomment it in prpclient ini ( and make sure you have in that directory all required DLL ( 32 or 64 bit) for cuda
____________
92*10^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)
4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME
2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT
Proud member of team Aggie The Pew. Go Aggie!

OK, so I got the latest version of llrCUDA compiled with CUDA 5.0, and got it running. It works with small K's:

3*2^382449+1 is prime! Time : 200.866 sec.
3*2^414840-1 is prime! Time : 220.356 sec.

But for larger K's I get:

Not support zero-padded rational base DWT

:|

So, is there a PRPNet port that does really small K's? Like a 321 port, say?

Edit: I tried primes from the 27121 project. Neither K is small enough. :(
____________

Curious if you've tried K=1 (yes I know there are already a whole bunch of people running that elsewhere). Just wondering if it works, and if so, how fast.

--Gary