LL with OpenCL - Page 9

Robish · 2013-09-07, 12:10

Fantastic works guys! well done! Ill try it now with a 7870 if I can.

cheers

Rob

msft · 2013-09-08, 08:59

Quote:

Originally Posted by Robish

Fantastic works guys! well done! Ill try it now with a 7870 if I can.

Please wait until the windows version available.

kracker · 2013-09-08, 13:36

Quote:

Originally Posted by msft

Please wait until the windows version available.

Still some loose ends on my side though.

Quote:

Originally Posted by msft

Normal.
It is slow except power of two.

So isn't it possible to not select fft lengths the power of two unless explicitly asked with -f?

msft · 2013-09-08, 14:38

Quote:

Originally Posted by kracker

So isn't it possible to not select fft lengths the power of two unless explicitly asked with -f?

This version Normalize can not support only pow2.
Too large fft size make error.

Code:

$ ./clLucas 1398269

start M1398269 fft length = 73728
Iteration 10000 M( 1398269 )C, 0xa4a6d2f0e34629db, n = 73728, clLucas v1.00 err = 0.09375 (1:01 real, 6.0343 ms/iter, ETA 2:18:47)

$ ./clLucas -f 147456 1398269

start M1398269 fft length = 147456
err = 16,fft length = 147456 exiting.
Warning:  Program terminating, but clFFT resources not freed.
Please consider explicitly calling clfftTeardown( ).

kracker · 2013-09-10, 01:14

M( 1398269 )P, n = 73728, CUDALucas v1.66

msft · 2013-09-10, 02:59

Quote:

Originally Posted by kracker

M( 1398269 )P, n = 73728, CUDALucas v1.66

Congratulations.
Thank you lots of work.

msft · 2013-09-10, 05:18

Quote:

Originally Posted by kracker

M( 1398269 )P, n = 73728, CUDALucas v1.66

I think your 7770 25% faster than my 7750.

LaurV · 2013-09-10, 07:55

Quote:

Originally Posted by msft

I think your 7770 25% faster than my 7750.

Someone post a win64 exe and I tell you how fast a 7970 and a 7990 are...

Huh? What do you think?

msft · 2013-09-10, 08:31

Quote:

Originally Posted by LaurV

Someone post a win64 exe and I tell you how fast a 7970 and a 7990 are...

Huh? What do you think?

It's can't be tell,Obesity.
400~800%,I guess.

kracker · 2013-09-10, 14:24

7750 vs two 7970.
http://anandtech.com/bench/product/535?vs=588

Also, DP ratio is diffrent, 1/16(?) for 7750, 1/4 for 79xx.

Right now doing a DC check with my 7770

TeknoHog · 2013-09-10, 17:52

I have followed this project with great interest, and I look forward to a usable release. So far, it has been a little hard to piece together all the information necessary for compilation. For example, when trying to compile 0.59, I was left wondering what openclsdkdefs.mk is supposed to be, and where I should find SDKApplication.hpp. I use Linux and I'm not exactly a beginner with these things, but for some reason I feel completely lost here, as if I was missing some essential package (besides AMD APP SDK and clAmdFft, of course).

One reason why I mention this now is that I recently released an automatic work assignment/submission tool for mfakto, and I'd like to extend it for LL testing. Of course, I would try CudaLucas if I had any Nvidia hardware.

2013-09-07, 12:10	#89
Robish "Rob Gahan" Aug 2013 Ireland 44₈ Posts	Great work Fantastic works guys! well done! Ill try it now with a 7870 if I can. cheers Rob

2013-09-10, 01:14	#93
kracker "Mr. Meeseeks" Jan 2012 California, USA 37×59 Posts	M( 1398269 )P, n = 73728, CUDALucas v1.66 Attached Thumbnails Last fiddled with by kracker on 2013-09-10 at 01:16

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
mfakto: an OpenCL program for Mersenne prefactoring	Bdot	GPU Computing	1724	2023-06-04 23:31
Can't get OpenCL to work on HD7950 Ubuntu 14.04.5 LTS	VictordeHolland	Linux	4	2018-04-11 13:44
OpenCL accellerated lattice siever	pstach	Factoring	1	2014-05-23 01:03
OpenCL for FPGAs	TObject	GPU Computing	2	2013-10-12 21:09
AMD's Graphics Core Next- a reason to accelerate towards OpenCL?	Belteshazzar	GPU Computing	19	2012-03-07 18:58

2013-09-10, 14:24	#98
kracker "Mr. Meeseeks" Jan 2012 California, USA 4207₈ Posts	7750 vs two 7970. http://anandtech.com/bench/product/535?vs=588 Also, DP ratio is diffrent, 1/16(?) for 7750, 1/4 for 79xx. Right now doing a DC check with my 7770

2013-09-10, 17:52	#99
TeknoHog Mar 2010 Jyvaskyla, Finland 44₈ Posts	I have followed this project with great interest, and I look forward to a usable release. So far, it has been a little hard to piece together all the information necessary for compilation. For example, when trying to compile 0.59, I was left wondering what openclsdkdefs.mk is supposed to be, and where I should find SDKApplication.hpp. I use Linux and I'm not exactly a beginner with these things, but for some reason I feel completely lost here, as if I was missing some essential package (besides AMD APP SDK and clAmdFft, of course). One reason why I mention this now is that I recently released an automatic work assignment/submission tool for mfakto, and I'd like to extend it for LL testing. Of course, I would try CudaLucas if I had any Nvidia hardware.