Sieving for CRUS

[rebirther]

Originally Posted by pepi37

New sieve batch will be sent to YOYO in few days. If some need some sieve, now is time :)
You can leave post here or on PM

Do we have already the rest of all recommended bases sieve files in pipeline?

[pepi37]

Originally Posted by rebirther

Do we have already the rest of all recommended bases sieve files in pipeline?

No , but we will have it: since nobody ask nothing: rest of recommended bases is priority one!

[pepi37]

r_602
r_643
r_667
r_694
r_696
r_803
r_887
r_1003
sent for sieving

[pepi37]

New set of data will be sent to Yoyo!
Does anyone has any "sieve wish"? :)

[gd_barnes]

Originally Posted by pepi37

New set of data will be sent to Yoyo!
Does anyone has any "sieve wish"? :)

R107, S368, S994 :-)

[rebirther]

R15 (10M-20M) n=25-100k

[pepi37]

Today I and YOYO will try our deepest sieve file ( and as such) we will pass our previous record of 20000 WU per sieve files. I test in on my two repdigit sequences. So if all be ok, new sieve depth can be 15e16


This will be used only on NPG files ( base with one K remain) since in that case one workunit is 500e9.


I will inform you in progress. Best regards to all

[pepi37]

Originally Posted by sweety439

You can sieve the 1k bases at n<1M with no sieve file to 1M: (see https://mersenneforum.org/showthread.php?t=12980), they are:

S117 500K-1M
S257 600K-1M
S430 500K-1M
S497 500K-1M
S789 500K-1M
S883 500K-1M
R883 500K-1M
R968 500K-1M

I can always sent those base for sieving, but until now I was limited to 1e15. So we now extended , but easy ,step by step.We can be sure YOYO system can handle without problem 40000 workunits per sequence

[gd_barnes]

Originally Posted by pepi37

I can always sent those base for sieving, but until now I was limited to 1e15. So we now extended , but easy ,step by step.We can be sure YOYO system can handle without problem 40000 workunits per sequence

Please do not listen to Sweety. He trolls here constantly and is not part of this project. He does not understand anything about work effort required because he never does any work for the project. The sieve depth that you are discussing is great but it would be much better done for sequences where n>1M. It would be a waste of sieving effort for n<1M.

One of our main goals for 2021 will be testing all bases with 9 or 10 k's remaining to n=200K. So if you can pick out some of those bases that are only tested to n=100K and sieve them for n=100K-500K that would be more helpful for the project.

In the next few days I will be posting our goals for 2021.

[KEP]

Just a lot of rambeling, did not understand that there was no need for initial sieving.

Happy new year

[pepi37]

Originally Posted by KEP

Hi

I've come to a point, where P95 and LLR work is no longer possible on my i5 4670, but it shows no signs of trouble when sieving and since sieving is less stressfull on the CPU, I'm going to retire the i5 4670 from LLR and devote it to sieving. In regards to my reservation on SR383 it will be completed, since I'm going to build 2 systems with Ryzen 9 3900X processor, wich will start out by doing about 1000 tests a day on S383 - so what would have been a delay has now forced me to upgrade to a scale where 20 times as much work can be done each day and to an extent such that maybe 2 primes for S383 is possible next year, even if I double check some or all of my completed work for S383 (work not already double checked by PG).

Now back to the offer of sieving, I have 4 cores, so can you give me 4 bases to sieve for n=100K to n=500K or to n=1M(?)?

I'm very late for this party, but are these following bases tested to n100K and with 9 k's remaining available for sieving:

S572 to n=500K or n=1M?
S639 to n=500K or n=1M?
S908 to n=500K or n=1M?
S977 to n=500K or n=1M?

How deep do I have to sieve, before sending to Pepi and Pepi sending to YOYO?

Thanks for any guidelines and if those above is not available please give me 4 different bases

Take care

KEP you dont need to donate your CPU for sieving: I made all: initial sieve before I sent to YOYO is done in minutes, and then it is sent to YOYO to finish. So use your CPU for you : every processed workunit is one less :)
Thanks for offering your computer power :)