Bases 33-100 reservations/statuses/primes - Page 157

Grotex · 2022-05-20, 05:39

R71 to k=1M,n=2500 done.

VBCurtis · 2022-05-24, 19:23

R36 is complete to 420k (840k base 6, as LLR expresses the candidates).
I've moved this job from an old Ivy Bridge to a new Ryzen 5950X, still two cores. I'm continuing for a while still, 500k at least.

MisterBitcoin · 2022-05-25, 19:04

Quote:

Originally Posted by MisterBitcoin

Reached n=771K, no primes found, extending up to n=900K.

S49 reached n=1M, no prime found. Results attached, releasing base.

Is anyone interested in takeing R71 k<1M? I would be interested to test it up to n=5K which should be possible in a short time; however up to n=10K is possibly better for BOINC. Otherwise I would need to spilt the base into smaller potions so that every core gets some work. (Lesson from R1012)

gd_barnes · 2022-05-25, 19:38

Quote:

Originally Posted by MisterBitcoin

however up to n=10K is possibly better for BOINC.

BOINC cannot reasonably do such small tests.

VBCurtis · 2022-05-25, 20:24

I think he meant that going to 10k himself would yield something whose next step is better for BOINC. The 5k-10k tests are a little small for BOINC, as you observe. That is, I think you two agree.

rogue · 2022-05-25, 20:56

PRPNet is also an option, but only if one is running on a local server. I have used it for S3 and R3.

MisterBitcoin · 2022-05-26, 00:23

Quote:

Originally Posted by rogue

PRPNet is also an option, but only if one is running on a local server. I have used it for S3 and R3.

Yes, that sounds doable however I never used that. Though I dont think it will help us right now with "just" an 1M range. I should look in that to speed up R1012.

And Gary is correct the tests are way too quick for BOINC.

gd_barnes · 2022-05-26, 06:49

Quote:

Originally Posted by VBCurtis

I think he meant that going to 10k himself would yield something whose next step is better for BOINC. The 5k-10k tests are a little small for BOINC, as you observe. That is, I think you two agree.

That isn't the way I read it. But regardless...even n=10k-25k tests are likely too small for BOINC.

The smallest that we've had them do is n=25k-100k for many bases > 100. Although the n=50k-100k we had them do for R3 were likely even faster tests. I know those were quite a bit of work for Reb.

pepi37 · 2022-06-04, 06:20

K4 S53 at 2.075 M

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Bases 251-500 reservations/statuses/primes	gd_barnes	Conjectures 'R Us	2437	2022-06-26 08:45
Bases 101-250 reservations/statuses/primes	gd_barnes	Conjectures 'R Us	989	2022-06-26 08:35
Riesel base 3 reservations/statuses/primes	KEP	Conjectures 'R Us	1134	2022-06-24 20:28
Bases 501-1030 reservations/statuses/primes	KEP	Conjectures 'R Us	4097	2022-06-20 09:23
Bases 6-32 reservations/statuses/primes	gd_barnes	Conjectures 'R Us	1417	2022-05-25 19:33

2022-05-24, 19:23	#1718
VBCurtis "Curtis" Feb 2005 Riverside, CA 5322₁₀ Posts	R36 is complete to 420k (840k base 6, as LLR expresses the candidates). I've moved this job from an old Ivy Bridge to a new Ryzen 5950X, still two cores. I'm continuing for a while still, 500k at least.

2022-05-25, 20:24	#1721
VBCurtis "Curtis" Feb 2005 Riverside, CA 14CA₁₆ Posts	I think he meant that going to 10k himself would yield something whose next step is better for BOINC. The 5k-10k tests are a little small for BOINC, as you observe. That is, I think you two agree.

2022-05-25, 20:56	#1722
rogue "Mark" Apr 2003 Between here and the 3·2,213 Posts	PRPNet is also an option, but only if one is running on a local server. I have used it for S3 and R3.

2022-06-04, 06:20	#1725
pepi37 Dec 2011 After milion nines:) 2²×383 Posts	K4 S53 at 2.075 M