A computation worthy of the name - Page 2

fivemack · 2009-09-10, 19:44

That looks like the right sort of timescale; I didn't expect it to make much of a difference, and the numbers are coming out as about a third of what I was getting on an otherwise-idle 64-bit C2/2400.

R.D. Silverman · 2009-09-11, 12:36

Quote:

Originally Posted by fivemack

I have spent a little while figuring out parameter choice for this; here is the polynomial for 2^941 - 1:

You may want to reserve this number with Sam.

fivemack: did this yesterday, have received a confirmation email but it's not in the list on the Web yet

bdodson · 2009-09-11, 14:53

On the "how much memory" thread that preceded this one, I reported

Quote:

PS - ecm is at 13.5*t50, mostly with p55- or p60-limits. That's
past 2*t55; maybe about half of t60? (c. = 25t50). I'd consider
a bit more, if we're serious, ...

I added another 4t50 (p60-limits), for a total of 17.5*t50. That's
something like 3*t55, not enough to regard a p59/p60 as a miss,
but enough to have had a fair shot at the rest of p5x < p59. And
that was before finding out just how serious we were!

-Bruce (at least 16t50 was with p60-limits, with p55 a bit past 5t50).

R.D. Silverman · 2009-09-11, 15:00

Quote:

Originally Posted by bdodson

And
that was before finding out just how serious we were!

Absolutely! Let us all B. cereus.

CRGreathouse · 2009-09-11, 17:42

Quote:

Originally Posted by R.D. Silverman

Absolutely! Let us all B. cereus.

I could find a Cereus b. but not a B. cereus.

fivemack · 2009-09-11, 17:43

Quote:

Originally Posted by CRGreathouse

I could find a Cereus b. but not a B. cereus.

http://en.wikipedia.org/wiki/Bacillus_cereus

CRGreathouse · 2009-09-11, 18:15

Quote:

Originally Posted by fivemack

http://en.wikipedia.org/wiki/Bacillus_cereus

I'm awed by your ability to find useless information. (I'll stop thread-jacking now.)

J.F. · 2009-09-15, 20:28

I'm having trouble finding a 16e binary... Would someone help me out?

fivemack · 2009-09-15, 22:27

http://www.chiark.greenend.org.uk/~t...sm64latest.zip

has the full set from 11 to 16

Batalov · 2009-09-16, 01:42

We will have to slightly update that file, because the binaries in it are dated Apr/15 ("the tax day special edition"), but later in the process of preliminary sieiving for the 5,398+ two new infinite loops were found and ad hoc patches were applied, around Jun/10/2009. All of them are in the SVN tree. (Oh, no, I haven't added them; I remember now, I was waiting for the large-scale tests to finish at Lehigh.edu. I'll add these two patches as SVN revision 367, right now.)

So far the binaries that I use even in the dangerous region (around the power of two) didn't have any problems. Tom, let me email you a replacement ZIP file, so you could replace the chiark mirror file, ok?

fivemack · 2009-09-16, 08:24

http://www.chiark.greenend.org.uk/~t...updated16e.zip

has the new version of the 16e binary (dated June 10th) in it, and I'll put up one with the new versions of all the binaries when batalov sends it to me.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Possible response to "only uploading is lawsuit-worthy."	jasong	jasong	3	2012-12-27 16:40
New Computation	JohnFullspeed	Miscellaneous Math	8	2011-07-13 10:54
New Pi Computation Record	ldesnogu	Lounge	11	2010-01-07 14:42
Value of computation	fivemack	Lounge	0	2008-09-05 20:23
Saving computation in ECM	dave_dm	Factoring	8	2004-06-12 14:18

2009-09-10, 19:44	#12
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 2·7·461 Posts	That looks like the right sort of timescale; I didn't expect it to make much of a difference, and the numbers are coming out as about a third of what I was getting on an otherwise-idle 64-bit C2/2400.

2009-09-15, 20:28	#19
J.F. Jun 2008 2³·3² Posts	I'm having trouble finding a 16e binary... Would someone help me out?

2009-09-15, 22:27	#20
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 1100100110110₂ Posts	http://www.chiark.greenend.org.uk/~t...sm64latest.zip has the full set from 11 to 16

2009-09-16, 01:42	#21
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 3²·1,117 Posts	We will have to slightly update that file, because the binaries in it are dated Apr/15 ("the tax day special edition"), but later in the process of preliminary sieiving for the 5,398+ two new infinite loops were found and ad hoc patches were applied, around Jun/10/2009. All of them are in the SVN tree. (Oh, no, I haven't added them; I remember now, I was waiting for the large-scale tests to finish at Lehigh.edu. I'll add these two patches as SVN revision 367, right now.) So far the binaries that I use even in the dangerous region (around the power of two) didn't have any problems. Tom, let me email you a replacement ZIP file, so you could replace the chiark mirror file, ok?

2009-09-16, 08:24	#22
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 6454₁₀ Posts	http://www.chiark.greenend.org.uk/~t...updated16e.zip has the new version of the 16e binary (dated June 10th) in it, and I'll put up one with the new versions of all the binaries when batalov sends it to me.