GPU LL Testing FAQ - Page 9

flashjh · 2012-03-08, 19:34

Quote:

Originally Posted by aaronhaviland

I started using one of these to move the middle card of my 3-GPU system for better airflow: http://www.amazon.com/gp/product/B0058UVVX2/
I might use another, to make it a 4-GPU system (it's a 4-slot mobo, in a too-small tower case)

Can you post a picture so I can see how it works with final setup? What PS do you use on that system?

Dubslow · 2012-04-06, 20:08

Whoa dang, the PDF is way out of date, we're up to CuLu 2.0 now, not 1.4.2 ;)

(Also, GMP-ECM has been ported at least sort of to the GPU, that should be removed.)

Brain · 2012-04-07, 10:07

I will probably do this soon. It's kind of useless to update the pdf when CL changes versions so quickly as it did the last weeks.
I monitor the CL thread and will release the update when I think that CL2 is ready for production work. I guess I will be convinced in the next few days.
Additionally, I dislike old versions wandering around in the stickies. Two options: Stop publishing the guide in the sticky and use a new thread or get mod rights. I prefer the former.
I am not convinced that CUDA-ECM should be integrated yet as it is expert and development level at the moment.
Thoughts? Is it useful at all?

Dubslow · 2012-04-07, 17:46

I wasn't saying it should be integrated, just remove the sentence saying it needs to be done.

xilman · 2012-04-08, 07:18

Quote:

Originally Posted by Brain

I am not convinced that CUDA-ECM should be integrated yet as it is expert and development level at the moment.
Thoughts? Is it useful at all?

IMO, it is very useful. I've already found three non-trivial factors with it and another day-long run is about 80% completed at the moment.

Paul

debrouxl · 2012-04-08, 08:57

Yeah, I've found some small factors as well, and I see that the ECMNET records page ( http://www.loria.fr/~zimmerma/records/ecmnet.html ) shows several factors whose stage 1 was performed by GPU-ECM (A = ... instead of sigma, and Cyril Bouvier being the discoverer).

My low-end GT 540M running GPU-ECM beta beats the associated Core i7-2670QM running GMP-ECM by a wide margin for some numbers, and other persons' tests show that the ratio is even more damning for CPUs if high-end GT(X) 4xx and 5xx are used.

xilman · 2012-04-08, 10:57

Quote:

Originally Posted by debrouxl

Yeah, I've found some small factors as well, and I see that the ECMNET records page ( http://www.loria.fr/~zimmerma/records/ecmnet.html ) shows several factors whose stage 1 was performed by GPU-ECM (A = ... instead of sigma, and Cyril Bouvier being the discoverer

That might be the case but I'm far from certain. Try running gmp-ecm with the -batch parameter to see what I mean.

Paul

Brain · 2012-04-08, 17:26

Quote:

Originally Posted by xilman

IMO, it is very useful. I've already found three non-trivial factors with it and another day-long run is about 80% completed at the moment.

Paul

My fault: The "Is it useful" was referring to the pdf guide. I will mention GMP-ECM thread in the next version. Just waiting for more news from mfakto - new version likely to come soon.

ixfd64 · 2012-07-13, 21:31

This is probably a stupid question, but would running multiple LL tests on one GPU yield a higher throughput?

For example, the GeForce GTX 570 has 480 CUDA cores, and I imagine that a lot of threads would get in each other's way. In this case, would it be better to run, say, ten parallel tests using 48 cores each, instead of doing one test with all 480 cores? I don't think memory bandwidth would be an issue as GPUs normally have much higher bandwidths than CPUs.

Dubslow · 2012-07-13, 21:41

I think the problem is that much of the code is actually CPU based, but it's lightweight-enough code that CUDALucas doesn't appear to use any CPU time. The point is, a lot of transfers between host and device occur each iteration, and adding any more transfers would significantly slow both tests running side by side. Furthermore, it simply isn't possible to run the test only on the GPU.

You'll have to wait for msft to verify/clarify these statements, since he's the only one (plus George) to write the math code, but this should provide at least a basic answer.

henryzz · 2012-07-13, 21:52

Quote:

Originally Posted by Dubslow

I think the problem is that much of the code is actually CPU based, but it's lightweight-enough code that CUDALucas doesn't appear to use any CPU time. The point is, a lot of transfers between host and device occur each iteration, and adding any more transfers would significantly slow both tests running side by side. Furthermore, it simply isn't possible to run the test only on the GPU.

You'll have to wait for msft to verify/clarify these statements, since he's the only one (plus George) to write the math code, but this should provide at least a basic answer.

Wouldn't this mean that if we got some sort of GPU LL test running on a ivy bridge it would run really fast?

2012-04-07, 10:07	#91
Brain Dec 2009 Peine, Germany 331₁₀ Posts	GPU Computing Guide Update I will probably do this soon. It's kind of useless to update the pdf when CL changes versions so quickly as it did the last weeks. I monitor the CL thread and will release the update when I think that CL2 is ready for production work. I guess I will be convinced in the next few days. Additionally, I dislike old versions wandering around in the stickies. Two options: Stop publishing the guide in the sticky and use a new thread or get mod rights. I prefer the former. I am not convinced that CUDA-ECM should be integrated yet as it is expert and development level at the moment. Thoughts? Is it useful at all? Last fiddled with by Brain on 2012-04-07 at 10:08

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Anti-poverty drug testing vs "high" tax deduction testing	kladner	Soap Box	3	2016-10-14 18:43
What am I testing?	GARYP166	Information & Answers	9	2009-02-18 22:41
k=243 testing ??	gd_barnes	Riesel Prime Search	20	2007-11-08 21:13
Testing	grobie	Marin's Mersenne-aries	1	2006-05-15 12:26
Speed of P-1 testing vs. Trial Factoring testing	eepiccolo	Math	6	2006-03-28 20:53

2012-04-06, 20:08	#90
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3·29·83 Posts	Whoa dang, the PDF is way out of date, we're up to CuLu 2.0 now, not 1.4.2 ;) (Also, GMP-ECM has been ported at least sort of to the GPU, that should be removed.)

2012-04-07, 17:46	#92
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3×29×83 Posts	I wasn't saying it should be integrated, just remove the sentence saying it needs to be done.

2012-04-08, 08:57	#94
debrouxl Sep 2009 2²·5·7² Posts	Yeah, I've found some small factors as well, and I see that the ECMNET records page ( http://www.loria.fr/~zimmerma/records/ecmnet.html ) shows several factors whose stage 1 was performed by GPU-ECM (A = ... instead of sigma, and Cyril Bouvier being the discoverer). My low-end GT 540M running GPU-ECM beta beats the associated Core i7-2670QM running GMP-ECM by a wide margin for some numbers, and other persons' tests show that the ratio is even more damning for CPUs if high-end GT(X) 4xx and 5xx are used.

2012-07-13, 21:31	#97
ixfd64 Bemusing Prompter "Danny" Dec 2002 California 2³×313 Posts	This is probably a stupid question, but would running multiple LL tests on one GPU yield a higher throughput? For example, the GeForce GTX 570 has 480 CUDA cores, and I imagine that a lot of threads would get in each other's way. In this case, would it be better to run, say, ten parallel tests using 48 cores each, instead of doing one test with all 480 cores? I don't think memory bandwidth would be an issue as GPUs normally have much higher bandwidths than CPUs.

2012-07-13, 21:41	#98
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3·29·83 Posts	I think the problem is that much of the code is actually CPU based, but it's lightweight-enough code that CUDALucas doesn't appear to use any CPU time. The point is, a lot of transfers between host and device occur each iteration, and adding any more transfers would significantly slow both tests running side by side. Furthermore, it simply isn't possible to run the test only on the GPU. You'll have to wait for msft to verify/clarify these statements, since he's the only one (plus George) to write the math code, but this should provide at least a basic answer.