mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   No Prime Left Behind (https://www.mersenneforum.org/forumdisplay.php?f=82)
-   -   Sieving drive for k=1005-2000 n=200K-500K (https://www.mersenneforum.org/showthread.php?t=11204)

gd_barnes 2009-01-14 01:13

[quote=nuggetprime;157884]8200G-8600G complete.9616 factors found. They will get mailed to Gary with in the next fifteen minutes.

nugget[/quote]


I PM'd Nugget yesterday about his factors. I still don't have them and he hasn't been on Mersenneforum since Jan. 10th. I'll try Emailing him now. If I don't hear from him by late Weds., I will start running his range myself. I think 2 quads will finish it up by Jan. 16th.

On another note, my 2 quads won't complete the P=8.8T-10T range until ~6 AM my time in the morning. So I'll move those 8 cores over to port 9000 sometime after that. But if I don't have Nugget's factors by late Weds. night, I'll have to move them back to sieving that range.


Gary

gd_barnes 2009-01-14 01:15

I've dropped the ball on computing an optimal sieving depth and don't have the time to mess with it right now. I think we can all agree that P=10T is enough. A rough estimate earlier had the optimum in the P=9T-10T range so this should be sufficient.

gd_barnes 2009-01-14 12:14

P=8800G-10000G is complete. no primes found

many possible primes were eliminated though!

IronBits 2009-01-14 12:39

Scrub a dub dub :thumbs:

IronBits 2009-01-14 13:43

[FONT=Verdana][COLOR=#0000ff]sieve 1005-2000-200K-500K done, factors emailed to Gary.

[/COLOR][/FONT]

em99010pepe 2009-01-14 13:50

[quote=IronBits;158699][FONT=Verdana][COLOR=#0000ff]sieve 1005-2000-200K-500K done, factors emailed to Gary.

[/COLOR][/FONT][/quote]

Now you can test your Core i7 on IB9000.

Mini-Geek 2009-01-14 14:33

[quote=gd_barnes;158686]P=8800G-10000G is complete. [B]no primes found[/B][/quote]
Naw, really? :razz:

gd_barnes 2009-01-15 05:27

[quote=Mini-Geek;158706]Naw, really? :razz:[/quote]


I was wondering if anyone would catch that.

Funny thing...I actually started to type that out of habit and then said what the heck am I doing? Then I just went with it. lol

If anyone finds a prime n>100K while sieving, let me know. We'll award you a gold medal! I say n>100K because srsieve will actually "announce" small primes on the screen if you sieve starting at an extremely low n-value like n=1.

gd_barnes 2009-01-15 06:14

Re-reserving Nugget's range of P=8200G-8600G. We can't wait any longer for it. Perhaps he'll show up midway through and I'll be able to stop. I think a quad will slay it fast enough. If not, I'll put 2 on it.


Gary

Mini-Geek 2009-01-15 16:31

[quote=gd_barnes;158779]I say n>100K because srsieve will actually "announce" small primes on the screen if you sieve starting at an extremely low n-value like n=1.[/quote]
I assume this is just because it sieves past the square root of the number without finding anything, right? How long would it take (I know it'll be way, way past the realm of practicality, and probably possibility as well) to confirm via sieving only, with sr*sieve or NewPGen or something, (maybe include Prime95?) that a number with n>100K is prime?

henryzz 2009-01-15 16:54

[quote=Mini-Geek;158819]I assume this is just because it sieves past the square root of the number without finding anything, right? How long would it take (I know it'll be way, way past the realm of practicality, and probably possibility as well) to confirm via sieving only, with sr*sieve or NewPGen or something, (maybe include Prime95?) that a number with n>100K is prime?[/quote]
almost right
it doesnt do square root for some reason although it would be very nice if it did

gd_barnes 2009-01-16 01:25

[quote=Mini-Geek;158819]I assume this is just because it sieves past the square root of the number without finding anything, right? How long would it take (I know it'll be way, way past the realm of practicality, and probably possibility as well) to confirm via sieving only, with sr*sieve or NewPGen or something, (maybe include Prime95?) that a number with n>100K is prime?[/quote]


Henry is correct. For example, if you are sieving and just happen to have k*2^n-1 calculate to a value of 101; srsieve will not display it as prime until it has sieved to P=101.

This is OK as far as I'm concerned because srsieve will actually correctly leave the pair in the sieve. sr1sieve and sr2sieve either barf at such a low pair or improperly remove them. NewPGen is extremely bad about that; it actually removes them from the sieve without saying a word. For that reason, I never use NewPGen anymore except for twin-prime sieving at higher n-ranges. Also, it's far too slow.


Gary

MyDogBuster 2009-01-16 08:32

4250G - 5850G Complete

Results emailed

I still owe you 4050G - 4250G Sometime tonight

MyDogBuster 2009-01-17 02:12

4050G - 4250G Complete

Results emailed

MyDogBuster 2009-01-17 02:30

In order to avoid a crunch on sieving, I'd think about starting a very very low intensity sieve of 1005-2000 n=500K-1MM. We will probably have to sieve to about 24-26T so getting started now would be wise.

gd_barnes 2009-01-17 03:21

[quote=MyDogBuster;159049]In order to avoid a crunch on sieving, I'd think about starting a very very low intensity sieve of 1005-2000 n=500K-1MM. We will probably have to sieve to about 24-26T so getting started now would be wise.[/quote]


Agreed totally. Actually, we thought far ahead on the n=600K-1M for k=400-1001 range and it was done several months before we got there but I totally misanticipated the speed at which the 5000th-place prime would rise to n>350K for the current range. PrimeGrid, searching in a style similar to our own for n=400K-500K on the Proth side, had a lot to do with that.

Now that you mention it, I'll start some sieving on one core for that within a couple of days and progress slowly. The low P-ranges take much longer and the file is very huge to start with but once it nears P=1T, then it's more manageable as a group effort.

This method of searching for primes is definitely the way to go in the future.


Gary

MyDogBuster 2009-01-17 03:34

[QUOTE]I'll start some sieving on one core for that within a couple of days and progress slowly. The low P-ranges take much longer and the file is very huge to start with but once it nears P=1T, then it's more manageable as a group effort.[/QUOTE]

Let me know if I can help. 1 core sounds about right here too.

gd_barnes 2009-01-17 08:09

[quote=MyDogBuster;159055]Let me know if I can help. 1 core sounds about right here too.[/quote]


OK, I'll take you up on that. I'll get it started and after I'm at about P=25G or so, we'll coordinate.

Thanks!

gd_barnes 2009-01-17 08:11

P=5.9T-7.1T is complete; now unreserved...factors Emailed.

MyDogBuster 2009-01-17 08:28

[QUOTE]P=5.9T-7.1T is complete; now unreserved...factors Emailed[/QUOTE]

Never tried emailing myself. Does it take very long? LOL

gd_barnes 2009-01-17 08:59

[quote=MyDogBuster;159065]Never tried emailing myself. Does it take very long? LOL[/quote]


No, just a few seconds.

That range is now available. Did you want to reserve it?

Glad to know folks are paying attention.

:missingteeth:

gd_barnes 2009-01-17 16:05

P=8.2T-8.6T is complete.

This sieving drive is now complete. Thanks to everyone for helping get this done quickly! :smile:

A file will be sent to David to load into port 8000 to restart the 8th drive here within the next half-hour.


Gary


All times are UTC. The time now is 04:37.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.