Quick Question about assignments

Dubslow · 2011-10-05, 22:58

51567343

This exponent last had TF activty in May 2010. I know me and many others have been getting hundreds and thousands of TF assignments up to 71 bits, but usually in the 54M-55M range, at least in my experience. How come this exponent hasn't been assigned for more TF'ing, and even if not, why did it take so long to get an LL assignment to it? I got 53M back in June, when this one was just as far factored as those. Why would this happen?

Uncwilly · 2011-10-05, 23:35

Maybe it had been picked up by an anon a while back and anon never reported in. Exponent then got reassigned to you.

Dubslow · 2011-10-06, 02:14

Possibly, but then why not TF instead of LL?

Prime95 · 2011-10-06, 03:42

Since GPUs have successfully got most exponents up to 2^70, I've started handing out some TF to 2^71 to keep the GPUs busy.

Dubslow · 2011-10-06, 03:47

But then why was this one sent straight to LL? It's only at 69.

Prime95 · 2011-10-06, 04:07

These extra bits of TF are done manually on a rather ad hoc basis.

Christenson · 2011-10-06, 04:19

and I repeat my offer to manually TF any exponent desired to a reasonable level on request.

I had another question:
I've just restarted CUDALucas, and fed it a 25M LL-D. Unlike working for P95, there doesn't seem to be any randomisation of the data passing through the busses, so if there were a data-dependent error, and I were to repeat the test, I would have a good chance of getting the same, erroneous residue.

How does Primenet ensure independence of two LL checks on the same exponent under such circumstances?

Prime95 · 2011-10-06, 08:46

Quote:

Originally Posted by Christenson

How does Primenet ensure independence of two LL checks on the same exponent under such circumstances?

It doesn't. You can't use CudaLUCAS to double-check an exponent that was first-time tested with CudaLUCAS.

Christenson · 2011-10-06, 11:57

Quote:

Originally Posted by Prime95

It doesn't. You can't use CudaLUCAS to double-check an exponent that was first-time tested with CudaLUCAS.

If primenet does not enforce the mathematically required independence, but I can manually asssign the same exponent to my GPU twice, how does that work?

Clearly, the concern is valid, especially since my GT480 is competitive with at least two CPU cores.

davieddy · 2011-10-06, 13:52

Quote:

Originally Posted by Christenson

If primenet does not enforce the mathematically required independence, but I can manually asssign the same exponent to my GPU twice, how does that work?

Clearly, the concern is valid, especially since my GT480 is competitive with at least two CPU cores.

Does CUDALucas employ the "random shift" trick described at the
bottom of "The Math" page?
It is trivial to implement it.

David

Prime95 · 2011-10-06, 14:05

Quote:

Originally Posted by Christenson

If primenet does not enforce the mathematically required independence, but I can manually asssign the same exponent to my GPU twice, how does that work?

Clearly, the concern is valid, especially since my GT480 is competitive with at least two CPU cores.

It is enforced when you check in the result. Two matching CudaLucas residues will not result in a change to the verified state.

2011-10-05, 22:58	#1
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3×29×83 Posts	Quick Question about assignments 51567343 This exponent last had TF activty in May 2010. I know me and many others have been getting hundreds and thousands of TF assignments up to 71 bits, but usually in the 54M-55M range, at least in my experience. How come this exponent hasn't been assigned for more TF'ing, and even if not, why did it take so long to get an LL assignment to it? I got 53M back in June, when this one was just as far factored as those. Why would this happen?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
A quick question	Pegos	Information & Answers	6	2016-08-11 14:39
Quick TF Question	Dubslow	GPU Computing	2	2011-10-27 04:49
Quick msieve question	alkirah	Msieve	2	2009-12-30 14:00
Quick question about P90 CPU metric	stars10250	PrimeNet	9	2008-08-31 23:58
Quick p-1 question	Unregistered	Software	8	2006-10-13 23:35

2011-10-05, 23:35	#2
Uncwilly 6809 > 6502 """"""""""""""""""" Aug 2003 101×103 Posts 2²·2,767 Posts	Maybe it had been picked up by an anon a while back and anon never reported in. Exponent then got reassigned to you.

2011-10-06, 02:14	#3
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 1110000110101₂ Posts	Possibly, but then why not TF instead of LL?

2011-10-06, 03:42	#4
Prime95 P90 years forever! Aug 2002 Yeehaw, FL 17·487 Posts	Since GPUs have successfully got most exponents up to 2^70, I've started handing out some TF to 2^71 to keep the GPUs busy.

2011-10-06, 03:47	#5
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 1110000110101₂ Posts	But then why was this one sent straight to LL? It's only at 69.

2011-10-06, 04:07	#6
Prime95 P90 years forever! Aug 2002 Yeehaw, FL 17·487 Posts	These extra bits of TF are done manually on a rather ad hoc basis.

2011-10-06, 04:19	#7
Christenson Dec 2010 Monticello 5·359 Posts	and I repeat my offer to manually TF any exponent desired to a reasonable level on request. I had another question: I've just restarted CUDALucas, and fed it a 25M LL-D. Unlike working for P95, there doesn't seem to be any randomisation of the data passing through the busses, so if there were a data-dependent error, and I were to repeat the test, I would have a good chance of getting the same, erroneous residue. How does Primenet ensure independence of two LL checks on the same exponent under such circumstances?