mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Conjectures 'R Us (https://www.mersenneforum.org/forumdisplay.php?f=81)
-   -   Sieving drive Riesel base 6 n=150K-1M (https://www.mersenneforum.org/showthread.php?t=12277)

gd_barnes 2009-08-14 05:40

Sieving drive Riesel base 6 n=150K-1M
 
[COLOR=black][FONT=Verdana]This is a sieving drive for the 8 k's remaining on Riesel base 6 for n=150K-1M. Karsten kindly provided his file that is sieved to P=0.8T as a starting point.[/FONT][/COLOR]

Sieving up to P=43T is needed for the n=500K-1M range for the 5 remaining k's.

Like before, sr2sieve is what we will use. It is recommended that you run 64-bit sr2sieve on a 64-bit machine. Let us know if you need the executable or more detailed instructions on using it. Here is an example of the command to execute at the command prompt:

sr2sieve -p 35e12 -P 40e12 -i sieve-riesel-base6-500K-1M.txt

The above would be if you were sieving P=35T-40T. The file is listed after the "-i" command and is the actual file name that is posted in a link below. Feel free to name it something shorter if you want or use the "srwork" older convention where you don't have to specify a file name.

When complete, you should have a factors.txt file. Just post the file here in this thread or if it is too big, please Email the file to me at:
gbarnes017 at gmail dot com

A P=1T range should take ~2 CPU days on a modern 64-bit machine. Please reserve ranges in multiples of P=1T and plan to reserve no more than ~7-10 days of work at a time. When making reservations, please post your estimated completion date. This can be seen in sr2sieve about one minute after you start your sieve.

Here is a link to the latest sieve file: [URL="http://www.noprimeleftbehind.net/sieve/sieve-riesel-base6-150K-1"][COLOR=#0066cc][URL]http://www.noprimeleftbehind.net/crus/sieve-riesel-base6-500K-1M.txt[/URL][/COLOR][/URL]. All factors up to P=43T have been removed. We will remove additional factors as the drive progresses to slightly speed up sieving.

Reservations:
[code]
P-range reserved by status est. completion date
0T-0.8T kar_bon complete
0.8T-32T Lennart complete
32T-35T gd_barnes complete
35T-39T Flatlander complete
39T-43T Batalov complete
[/code]

[B]The sieving drive is complete![/B]

This is a very exciting base that can easily be proven within the next few years. All help is greatly appreciated! :smile:


Thank you,
Gary

gd_barnes 2009-08-14 09:41

Important note about the sieve file: Although it is included in the file, k=1597 was only sieved to n=500K. If anyone would like to take on the task of sieving just this one k for n=500K-1M up to our current sieving limit for the drive, let me know and we'll keep a non-formal posting about what ranges are being done and who is doing them if more than one person.


Gary

Lennart 2009-08-14 11:17

[quote=gd_barnes;185500]Important note about the sieve file: Although it is included in the file, k=1597 was only sieved to n=500K. If anyone would like to take on the task of sieving just this one k for n=500K-1M up to our current sieving limit for the drive, let me know and we'll keep a non-formal posting about what ranges are being done and who is doing them if more than one person.


Gary[/quote]


I take k 1597

Lennart

Lennart 2009-08-14 15:58

I have done k=1597 to 800G

I start sieve all from 800G-3T

Lennart

gd_barnes 2009-08-14 18:51

Thanks for the huge boost Lennart.

Lennart 2009-08-15 20:17

Sieved to 6T
 
1 Attachment(s)
Here is the file sieved to 6T


Lennart

gd_barnes 2009-08-18 01:35

Thanks a bunch Lennart!

To all:

In the next few hours, we'll start a drive for Riesel base 6 starting from n=150K on the final 8 k's remaining.

With such a low base and few k's remaining, this one has a very good possibility of being proven over the next few years. As a SWAG without doing any math or calculation, I would put its final prime in the vicinity of n=~10M.


Gary

mdettweiler 2009-08-18 05:33

[quote=gd_barnes;186172]As a SWAG without doing any math or calculation, I would put its final prime in the vicinity of n=~10M.[/quote]
YOWCH! :shock: Even with the new and improved PFGW, a test at that size for base 6 would probably take days to complete. Of course, that's not factoring in any expected increases in computer technology between now and when we get near that point.

gd_barnes 2009-08-18 06:30

[quote=mdettweiler;186187]YOWCH! :shock: Even with the new and improved PFGW, a test at that size for base 6 would probably take days to complete. Of course, that's not factoring in any expected increases in computer technology between now and when we get near that point.[/quote]

That is a very big SWAG and this particular base has a somewhat unique issue that makes it a little trickier than one might expect for as prime as it is. That of k=1597. See last para.

I agree with your comments but believe me, that is very good for most of the bases here. I think there's a reasonable chance that we could prove it in 10-20 years depending on how much we dedicate to it.

That's for the Riesel side but on the Sierp side, with 28 k's remaining, it likely will not be proven in any of our lifetimes even allowing for computer speed increases and improvements in software unless new mathematical methods are discovered for searching the bases. Even if it was base 2, it would likely be more difficult than the Riesel side.

People don't realize how hard it is to knock out the final k's on these bases. Even for a fairly prime base like base 6, it's tough because the final few k's are almost always very low weight.

Consider that there are only 5128 pairs remaining for the entire n=150K-1M range for k=1597 on a sieve to P=6T. k=1597 will be the key as it is by far the lowest weight k remaining. If we can pick up a prime for it for n<=1M, then there's a reasonable chance that we could find them all by n=3M to 5M. If not, well, this may end up being yet another base that we have one k remaining on for an extended period.

All of that said, I still consider base 6 to be very exciting because it is the only truly "low" base that is kind of difficult to prove but not so overwhelmingly so that one side can't be done in our lifetimes. In this case, I define "low" as < 8. All other bases < 8 will be terribly difficult to prove in many lifetimes (Sierp base 2 has a chance but still a somewhat low chance at this point since it's already at n>10M with 5 k's remaining). I believe these "somewhat difficult" bases are the most interesting. An example is the Dual Sierp base 2 project with 3 k's remaining at n=~3M. I think that one has a good chance of being proven by perhaps the n=20M to 30M range. In other words, it's tough but not too tough, the most interesting kind.


Gary

mdettweiler 2009-08-18 14:43

[quote=gd_barnes;186192]That is a very big SWAG and this particular base has a somewhat unique issue that makes it a little trickier than one might expect for as prime as it is. That of k=1597. See last para.

I agree with your comments but believe me, that is very good for most of the bases here. I think there's a reasonable chance that we could prove it in 10-20 years depending on how much we dedicate to it.

That's for the Riesel side but on the Sierp side, with 28 k's remaining, it likely will not be proven in any of our lifetimes even allowing for computer speed increases and improvements in software unless new mathematical methods are discovered for searching the bases. Even if it was base 2, it would likely be more difficult than the Riesel side.

People don't realize how hard it is to knock out the final k's on these bases. Even for a fairly prime base like base 6, it's tough because the final few k's are almost always very low weight.

Consider that there are only 5128 pairs remaining for the entire n=150K-1M range for k=1597 on a sieve to P=6T. k=1597 will be the key as it is by far the lowest weight k remaining. If we can pick up a prime for it for n<=1M, then there's a reasonable chance that we could find them all by n=3M to 5M. If not, well, this may end up being yet another base that we have one k remaining on for an extended period.

All of that said, I still consider base 6 to be very exciting because it is the only truly "low" base that is kind of difficult to prove but not so overwhelmingly so that one side can't be done in our lifetimes. In this case, I define "low" as < 8. All other bases < 8 will be terribly difficult to prove in many lifetimes (Sierp base 2 has a chance but still a somewhat low chance at this point since it's already at n>10M with 5 k's remaining). I believe these "somewhat difficult" bases are the most interesting. An example is the Dual Sierp base 2 project with 3 k's remaining at n=~3M. I think that one has a good chance of being proven by perhaps the n=20M to 30M range. In other words, it's tough but not too tough, the most interesting kind.


Gary[/quote]
Ah, that makes sense now. Thanks for the explanation! :smile:

gd_barnes 2009-09-29 08:47

The Riesel base 6 drive is closing in on completion to n=500K. Therefore we will need additional sieving for the n=500K-1M portion of the file. The optimum sieve depth for breaking off the n=500K-750K range is P=50T. We (mostly Lennart) have already sieved up to P=32T.

A P=1T range should take ~2 CPU days and we are asking that people reserve ranges in at least P=1T increments.

I am reserving P=32T-35T for one core. ETA is Oct. 5th.

Any and all help is greatly appreciated.


Thank you,
Gary

gd_barnes 2009-10-03 03:14

With the recent huge prime found on Riesel base 6, the optimum sieve depth should now be lower. When the prime is proven and submitted, I'll post the new depth here. My guess is that it will still come in somewhere between 40T and 45T.

Flatlander 2009-10-03 11:30

Taking 35-39T.

Flatlander 2009-10-03 12:31

[QUOTE=Flatlander;191779]Taking 35-39T.[/QUOTE]
ETA 7th Oct

gd_barnes 2009-10-03 14:42

I'll remove k=78959 from the posted sieve file later today. Chris, you may want to do that before starting. Thanks for helping out! :smile: If the optimum depth turns out to be P=45T, when my range finishes in a couple of days, I'll reserve the rest and put 2 cores on it to finish it up in ~5-6 days after that.

I'll post the new optimum depth when I update file later today.

Batalov 2009-10-03 21:03

I'll do 39-43T.
The latest sr2sieve? 1.8.11?
I am trying both 1.8.10 and .11. Both work well. ETA Oct 7
[FONT=Arial Narrow]39004844249663 | 51017*6^964687-1[/FONT]
[FONT=Arial Narrow]41003460163933 | 51017*6^862123-1[/FONT]
[FONT=Arial Narrow]...[/FONT]

mdettweiler 2009-10-03 23:22

[quote=Batalov;191809]I'll do 39-43T.
The latest sr2sieve? 1.8.11?
I am trying both 1.8.10 and .11. Both work well. ETA Oct 7
[FONT=Arial Narrow]39004844249663 | 51017*6^964687-1[/FONT]
[FONT=Arial Narrow]41003460163933 | 51017*6^862123-1[/FONT]
[FONT=Arial Narrow]...[/FONT][/quote]
As I recall, 1.8.10 is supposed to be a wee bit faster for sieves containing 5 or more sequences, so you should probably use that for this sieve.

Batalov 2009-10-04 00:06

The speeds are within 0.2-0.5%. In fact sometimes .11 is faster.

P.S. I think that posting a not-additionally-sieved file for the 500-550K range for PRP testing would be useful for someone.

gd_barnes 2009-10-04 06:10

The adjusted optimum sieve depth for breaking off n=500K-750K is P=42T. Calculation:

Testing time of an n=675K candidate (70% of 500K-750K n-range): 6282 secs.

Factor removal rate for the remaining 6 k's at P=32T: 4807 secs.

Optimum sieve depth: 6282 / 4807 * 32T = ~42T.

Therefore, all needed ranges are now reserved. Batalov, if you only want to do P=39T-42T instead of 39T-43T, that will be sufficient. Thanks for the extra boost.

I have now updated the sieve file in the 1st posting here to remove k=78959.


Gary

gd_barnes 2009-10-04 06:16

[quote=Batalov;191815]P.S. I think that posting a not-additionally-sieved file for the 500-550K range for PRP testing would be useful for someone.[/quote]

With k=78959 removed, all needed sieving ranges have been reserved and should be complete in < 4 days. Would you like to start on n=500K-550K with a 32T sieve depth now? If so, I can post it in the 1st posting here. Doing that would only be slightly less efficient than waiting on sieving up to 42T or 43T to finish.

The only thing we're asking is that people finish their ranges in < ~2 weeks. Keep in mind that n>500K ranges will contain all remaining 6 k's whereas the n=370K-500K range only contained 4 k's since Lennart had already completed 3 of them.


Gary

gd_barnes 2009-10-05 11:41

P=32T-35T is complete.

Batalov 2009-10-05 17:50

1 Attachment(s)
39-43T complete. 110 factors, expected 102.88.

Flatlander 2009-10-07 11:23

1 Attachment(s)
35T-39T is complete.

gd_barnes 2009-10-07 11:48

Great. The sieving drive is now complete for breaking off n=500K-750K. Later today, Max or I will apply the factors to the file and post a new file in team drive #4.

gd_barnes 2009-10-07 20:06

Max,

I went ahead and upload the file sieved to P=43T to the link in the first posting here. You can use it for matching up any results that you need to convert from PRPnet to LLR format.


Gary

mdettweiler 2009-10-08 01:16

[quote=gd_barnes;192136]Max,

I went ahead and upload the file sieved to P=43T to the link in the first posting here. You can use it for matching up any results that you need to convert from PRPnet to LLR format.


Gary[/quote]
Thanks, got it.

gd_barnes 2009-10-08 21:00

Max,

I have removed k=57023 from the posted file.


Gary

mdettweiler 2009-10-08 21:49

[quote=gd_barnes;192260]Max,

I have removed k=57023 from the posted file.


Gary[/quote]
Okay, got it.


All times are UTC. The time now is 05:37.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.