Bases 501-1030 reservations/statuses/primes - Page 134

f1pokerspeed · 2012-09-15, 21:46

Sorry for double post - reserving R894 - note that I am only taking k = 6. Will go as far as to prove it hopefully.

rogue · 2012-09-15, 21:59

Why don't you just continue with S991 and take it to n=50000? We're in no hurry.

f1pokerspeed · 2012-09-15, 22:13

In 15 days, I got 160 work units done. There were about 4,000 workunits in the initial batch - at that rate, I'd be done in a year. 25,000 to 50,000 would take about 4 years and two months.

That kind of timeframe is overdoing it for the first reservation I ever make. At the current rate i'm rolling now, I'll be at n=30k for k=6 in the next hour - n=100k in about 12 hours of wall clock time. That's a lot more achievable - I could do that this weekend.

gd_barnes · 2012-09-15, 23:04

Quote:

Originally Posted by f1pokerspeed

Sorry for double post - reserving R894 - note that I am only taking k = 6. Will go as far as to prove it hopefully.

Be sure and use the sieve file link on the Riesel reservations page. It is sieved for n=25K-100K to P=1T. You can quickly use the srfile program to remove all k's that are not = 6. Note that there are "only" 1074 tests for k=6 for this n-range and sieve depth. On a base of this size sieved that far for just one k, that gives you only about a 13-14% chance of finding a prime by n=100K. Therefore even testing all the way to n=100K where the tests can push a half hour each, the chances are small that you will find a prime.

Just figured I'd give you a heads up on the amount of CPU effort required on large bases here at CRUS.

gd_barnes · 2012-09-15, 23:21

Quote:

Originally Posted by f1pokerspeed

In 15 days, I got 160 work units done. There were about 4,000 workunits in the initial batch - at that rate, I'd be done in a year. 25,000 to 50,000 would take about 4 years and two months.

That kind of timeframe is overdoing it for the first reservation I ever make. At the current rate i'm rolling now, I'll be at n=30k for k=6 in the next hour - n=100k in about 12 hours of wall clock time. That's a lot more achievable - I could do that this weekend.

How many cores do you have? You might want to recheck those calculations for CPU time needed on k=6 for R894. See my previous post.

f1pokerspeed · 2012-09-16, 09:43

I have a quad-core laptop and a dual-core desktop. The laptop has AVX, but as a consequence of that it runs out of battery quickly.

I thought that a prime before n=100k might be possible

(NB: I extrapolated on the total testing time for a single core - oh well, it may be a couple of days then. Time to fire up the desktop...

gd_barnes · 2012-09-16, 21:23

Quote:

Originally Posted by f1pokerspeed

I have a quad-core laptop and a dual-core desktop. The laptop has AVX, but as a consequence of that it runs out of battery quickly.

I thought that a prime before n=100k might be possible

(NB: I extrapolated on the total testing time for a single core - oh well, it may be a couple of days then. Time to fire up the desktop...

I would suggest leaving the laptop plugged into an outlet and removing the battery. Then you have no worries about battery life. Also, the heat generated by the testing lowers the life of the battery quite a bit.

My estimate for k=6 n=25K-100K was around 15 CPU days. So yeah, 2-3 days if running 6 cores 24x7.

CGKIII · 2012-09-20, 01:29

Reserving S640 (new, ck = 11925) to n=25k.
Reserving S642 (new, ck = 10932) to n=25k.

MyDogBuster · 2012-09-22, 21:41

R653 tested n=25K-100K - nothing found

6k's remain - Results emailed

MyDogBuster · 2012-09-23, 18:08

Reserving S943 as new to n=25K

Mathew · 2012-09-24, 22:55

R1027 is complete to n=25K.

122 primes found. Results have been emailed.

2012-09-15, 22:13	#1466
f1pokerspeed Jun 2012 2×53 Posts	In 15 days, I got 160 work units done. There were about 4,000 workunits in the initial batch - at that rate, I'd be done in a year. 25,000 to 50,000 would take about 4 years and two months. That kind of timeframe is overdoing it for the first reservation I ever make. At the current rate i'm rolling now, I'll be at n=30k for k=6 in the next hour - n=100k in about 12 hours of wall clock time. That's a lot more achievable - I could do that this weekend. Last fiddled with by f1pokerspeed on 2012-09-15 at 22:14

2012-09-16, 09:43	#1469
f1pokerspeed Jun 2012 152₈ Posts	I have a quad-core laptop and a dual-core desktop. The laptop has AVX, but as a consequence of that it runs out of battery quickly. I thought that a prime before n=100k might be possible (NB: I extrapolated on the total testing time for a single core - oh well, it may be a couple of days then. Time to fire up the desktop... Last fiddled with by f1pokerspeed on 2012-09-16 at 09:46

2012-09-22, 21:41	#1472
MyDogBuster May 2008 Wilmington, DE 2²×23×31 Posts	R653 R653 tested n=25K-100K - nothing found 6k's remain - Results emailed Last fiddled with by MyDogBuster on 2012-09-22 at 21:41

2012-09-23, 18:08	#1473
MyDogBuster May 2008 Wilmington, DE 2²×23×31 Posts	S943 Reserving S943 as new to n=25K

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Riesel base 3 reservations/statuses/primes	KEP	Conjectures 'R Us	1106	2021-07-15 22:14
Bases 251-500 reservations/statuses/primes	gd_barnes	Conjectures 'R Us	2299	2021-07-11 08:19
Bases 101-250 reservations/statuses/primes	gd_barnes	Conjectures 'R Us	904	2021-07-06 18:51
Bases 33-100 reservations/statuses/primes	Siemelink	Conjectures 'R Us	1691	2021-07-06 18:50
Bases 6-32 reservations/statuses/primes	gd_barnes	Conjectures 'R Us	1394	2021-04-17 15:07

2012-09-15, 21:46	#1464
f1pokerspeed Jun 2012 2·53 Posts	Sorry for double post - reserving R894 - note that I am only taking k = 6. Will go as far as to prove it hopefully.

2012-09-15, 21:59	#1465
rogue "Mark" Apr 2003 Between here and the 11·577 Posts	Why don't you just continue with S991 and take it to n=50000? We're in no hurry.

2012-09-20, 01:29	#1471
CGKIII Aug 2012 5² Posts	Reserving S640 (new, ck = 11925) to n=25k. Reserving S642 (new, ck = 10932) to n=25k.

2012-09-24, 22:55	#1474
Mathew Nov 2009 2×5²×7 Posts	R1027 is complete to n=25K. 122 primes found. Results have been emailed.