Choose your own K and work on finding a top-5000 prime!

[VBCurtis]

Also reserving k=923. Looks like a very high-weight for a non-3*k value.
-Curtis

[gd_barnes]

Originally Posted by kar_bon

oh. no problem! i'm hunger for new informations to put them in the summary pages, to fill gaps and kill errors!!!
keep our summary for all k*2^n-1 the best, actual and completest on net is my purpose!
i'll keep an eye on this forum (and quite many other pages like Top5000) every day and all new info i obtain i put in the summary.
so sending your ultimate gap-filling file is like christmas and easter at one day for me

OK, you asked for it. I'll send my very big file your way later today.

But first, I've analyzed and checked all k's where 1000 < k < 10000 on the summary site that have primes listed where n <= 10000. This was a comprehensive check on the entire range, not just a spot check. I found 4 problems. (Quite excellent really!) After the corrections are made, I can personally guarantee that every prime shown that is n<=10000 will be correct for that range of k on the summary pages. I may do this for k < 1000 at some point also but I feel that the primes for k > 1000 would be far more prone to error because the smaller k's have had much more coordinated efforts. I seriously doubt that there are any errors on small n's on the k < 300 site. I anticipate maybe 1-2 errors for small n's for the Prime Search range of 300 < k < 1000.

I'll post the problems in the 'new data page' forum where you ask for error checking and that I've posted some problems before.


Gary

[arminius]

Reserving: 1000065.

First primes from 1 to 50k:

3, 5, 7, 14, 19, 20, 26, 81, 82, 113, 117, 146, 190, 268, 298, 308, 341, 391, 526, 979, 1122, 1221, 1310, 1454, 2146, 2468, 2564, 2938, 3047, 3119, 3791, 4395, 6303, 6439, 7117, 7635, 8165, 10082, 10357, 12532, 22597, 25263, 25411, 25713, 32663, 35845, 46340

[kar_bon]

to arminius:
1000065 is also prime for n=3, 5 and 7! (inserted in Summary)
Karsten

[arminius]

Ah - thank you!

[gd_barnes]

After initially doing a little testing on this one to fill some gaps, I realized that it is a HUGE performer for its size!

After completion of gap filling on it, I plan to include it in my arsenal of heavy-weights that I am sieving all at once from n=200K to 400K.


Gary

[gd_barnes]

Originally Posted by VBCurtis

Gary-- if you play with the sieving program a bit, you'll note that the sieve is roughly the same speed no matter how large a range of k you choose. I chose 1G because I was sieving on an old 128mb machine, and that was the biggest range that sieved at full speed. You may wish to reconsider doing the 100-500M block, if you even possibly might do another block later.

fixed-k sieves scale with the square root of n-range; fixed-n (including twin) sieves hardly scale at all.

-Curtis

Curtis,

This was from a while back so I thought I'd give you an update. Since we exchanged notes about us both doing a TPS in the n=90K to 110K range, I got access to another somewhat high-speed machine so have started up again on it. In doing so, I completed sieving on the k=100-500M block and have tested up to about k=200M on it. Since the odds are still well against me finding a twin in that range, last week, I started a sieve for the k=500M to 4G range. The machine has 1G memory so that isn't an issue. The main thing is that fortunately the range is just barely able to fit in NewPGen's max memory allocation of 485M so it is sieving at full speed. I calculated that testing the entire range of k=0 to 4G should give me about a 66% chance of a twin that would be in the top-10 (at the current time). Although total testing time on the one machine would be nearly 6 months if a twin isn't found sooner. I hope it won't take that long and more than that, I hope it won't take a bigger range! If so, I may have to pull a machine off of my heavy-weight k search and I don't really want to do that.


Gary

[gd_barnes]

A monster-weight for its size, I am reserving k=120023475 for inclusion in my huge sieve of high-weight k's from n=200K to 400K. I think this is the smallest k to have a Nash weight > 7000 shown on the summary site.

I'll fill the gap below prior testing and then separately test up to n=200K first.


Gary

[gd_barnes]

I'm going to reserve these smaller k's to begin a 'side effort' separate from my large-k very-heavy-weight search. It'll be nice to search some k's that don't take so long. I'll fill the large gaps in k=26565 and 49335 and then test them all from where they were left off at.


Gary

[gd_barnes]

Since I now have 16 k's reserved, I want to put all of them in one place and my planned upcoming efforts on them.

I am done sieving the range of n = 200K to 400K and am now LLRING all of my 12 large heavy-weight k's as follows:
19437
102765
3545685
111546435
115029915
120023475
290499495
686701125
775784295
968911515
1019340795
3428677395

1. 2 cores will be working on the range of n = 333335 to 400K to get a top-5000 prime quickly.
2. 1 core now and a 2 core later will be working on the range of n = 200K to 333334 to fill in the gap.

I also have reserved the following k's and after getting them tested and verified up to n = 200K, I will be doing the same with them as the above in a separate effort.
16995
26565
49335

I also have reserved k=2145. I get sieve files from Curtis periodically on that one and test them shortly afterword. It has now been tested up to n=280K.


Gary

reserving 28397655

currently sieving