The P-1 factoring CUDA program - Page 7

owftheevil · 2013-03-10, 16:39

Just a quick progress report. Stage 2 is coming along nicely. Referring to (p*k)^e - rp^e, it understands different values of e and p, I'm teaching it about rp now. Then its just a matter of choreography and logistics. The general multiplication kernel is functioning properly, a quick and dirty test on M51943429 with b1 = 1248, b2 = 1249, e = 2, and p = 6 found the factor.

ET_ · 2013-03-10, 17:34

Quote:

Originally Posted by owftheevil

Just a quick progress report. Stage 2 is coming along nicely. Referring to (p*k)^e - rp^e, it understands different values of e and p, I'm teaching it about rp now. Then its just a matter of choreography and logistics. The general multiplication kernel is functioning properly, a quick and dirty test on M51943429 with b1 = 1248, b2 = 1249, e = 2, and p = 6 found the factor.

swl551 · 2013-03-10, 18:06

I realize your work is still under development, but I have to ask if CudaP-1 will have a similar configuration/.INI to MFAKTO/C. Of course, I ask this to determine or sway its development to ensure MISFIT can easily work with the final product. Ideally MISFIT for MFAKTO/C would be able to front CudaP-1 with only minor changes.

Thanks

Scott

ET_ · 2013-03-10, 18:11

Quote:

Originally Posted by swl551

I realize your work is still under development, but I have to ask if CudaP-1 will have a similar configuration/.INI to MFAKTO/C. Of course, I ask this to determine or sway its development to ensure MISFIT can easily work with the final product. Ideally MISFIT for MFAKTO/C would be able to front CudaP-1 with only minor changes.

Thanks

Scott

CUDA P-1 is stil under development, and AFAIK (and have tested) there is still no user interface (though I found arecall to a !worktodo" file).

I expect that the final versio should be similar to either Mfakt? or CudaLucas to be usable with manual reservations.

Luigi

Dubslow · 2013-03-10, 23:56

The "final" version should be easy to make work the same way as CUDALucas, seeing as it's a CUDALucas conversion.

Stef42 · 2013-03-11, 15:59

I suppose there is only a linux version atm?

Dubslow · 2013-03-11, 16:30

Quote:

Originally Posted by Stef42

I suppose there is only a linux version atm?

flash/Brain ought to be able to compile it for Winbloze.

Stef42 · 2013-03-11, 18:01

That would be very nice. I got one GTX 590 and one GTX 570 running full speed, P-1 could be a nice addition to that

Brain · 2013-03-12, 21:17

Quote:

Originally Posted by Dubslow

flash/Brain ought to be able to compile it for Winbloze.

I cannot compile for Windows for at least 2 reasons:
- #include <gmp.h> fails. Changed to "gmp.h" but I have no Windows GMP library, only the header file gmp.h
- #include <cuda_safecalls.h> fails
I could make an investigation but don't have much time and I'm no linker expert. If anybody else can offer/point to precompiled libs and a suitable VC10 makefile...

Dubslow · 2013-03-12, 21:33

Hmm... I'm decently certain that the cuda_safecalls.h is included in CUDALucas as well, but I'm absolutely certain that the GMP dependency is new to P-1. You or others who desire a compilation ought to ask Brian Gladman for help; he's the forum's resident Windows/VS/GMP-for-Windows-aka-MPIR expert.

ET_ · 2013-03-21, 20:32

Question.

How is the GPU memory use computed, related to exponent, B1 and B2?

In other words, how can I know if B1 and B2 ranges fit in my GPU memory?

Luigi

2013-03-10, 16:39	#67
owftheevil "Carl Darby" Oct 2012 Spring Mountains, Nevada 3²×5×7 Posts	Just a quick progress report. Stage 2 is coming along nicely. Referring to (pk)^e - rp^e, it understands different values of e and p, I'm teaching it about rp now. Then its just a matter of choreography and logistics. The general multiplication kernel is functioning properly, a quick and dirty test on M51943429 with b1 = 1248, b2 = 1249, e = 2, and p = 6 found the factor. Last fiddled with by owftheevil on 2013-03-10 at 16:43 Reason: added the values for p and e*

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
mfaktc: a CUDA program for Mersenne prefactoring	TheJudger	GPU Computing	3628	2023-04-17 22:08
World's second-dumbest CUDA program	fivemack	Programming	112	2015-02-12 22:51
World's dumbest CUDA program?	xilman	Programming	1	2009-11-16 10:26
Factoring program need help	Citrix	Lone Mersenne Hunters	8	2005-09-16 02:31
Factoring program	ET_	Programming	3	2003-11-25 02:57

2013-03-10, 18:06	#69
swl551 Aug 2012 New Hampshire 328₁₆ Posts	I realize your work is still under development, but I have to ask if CudaP-1 will have a similar configuration/.INI to MFAKTO/C. Of course, I ask this to determine or sway its development to ensure MISFIT can easily work with the final product. Ideally MISFIT for MFAKTO/C would be able to front CudaP-1 with only minor changes. Thanks Scott

2013-03-10, 23:56	#71
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3×29×83 Posts	The "final" version should be easy to make work the same way as CUDALucas, seeing as it's a CUDALucas conversion.

2013-03-11, 15:59	#72
Stef42 Feb 2012 the Netherlands 2²·19 Posts	I suppose there is only a linux version atm?

2013-03-11, 18:01	#74
Stef42 Feb 2012 the Netherlands 4C₁₆ Posts	That would be very nice. I got one GTX 590 and one GTX 570 running full speed, P-1 could be a nice addition to that

2013-03-12, 21:33	#76
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3×29×83 Posts	Hmm... I'm decently certain that the cuda_safecalls.h is included in CUDALucas as well, but I'm absolutely certain that the GMP dependency is new to P-1. You or others who desire a compilation ought to ask Brian Gladman for help; he's the forum's resident Windows/VS/GMP-for-Windows-aka-MPIR expert.

2013-03-21, 20:32	#77
ET_ Banned "Luigi" Aug 2002 Team Italia 5·7·139 Posts	Question. How is the GPU memory use computed, related to exponent, B1 and B2? In other words, how can I know if B1 and B2 ranges fit in my GPU memory? Luigi