mersenneforum.org Reservations
 Register FAQ Search Today's Posts Mark Forums Read

2016-12-06, 17:07   #12
feromant

"Roman"
Dec 2016
Everywhere

2×13 Posts

Quote:
 mmff can't be used for such N and k, because of the bit depth of kernels.
mmff v0.28 handles divisors of size up to 252 bits. For the given range will be used by the kernels mfaktc_barrett224_F160_191gs and mfaktc_barrett236_F160_191gs.

2016-12-06, 17:17   #13
rogue

"Mark"
Apr 2003
Between here and the

172416 Posts

Quote:
 Originally Posted by ET_ Code: 10003-10999 50000000 269000000 5913 days with ppsieve_cuda & pfgw 14501-14999 30000000 70000000 640 days with ppsieve_cuda & pfgw 80001-80999 80000 100000 150 days with ppsieve_cuda & pfgw 114000-119999 40000 99999 12 years with ppsieve_cuda and pfgw 120001-129999 30000 40000 4 years with ppsieve_cuda & pfgw 130001-139999 20000 30000 5 years with ppsieve_cuda & pfgw 160001-169999 15000 20000 3 years with ppsieve_cuda & pfgw 180000 15000 30000 no data available 180001-181742 13000 30000 no data available 181744-189999 13000 30000 no data available
For these is there a limit for k where ppsieve is slower than fermfact?

2016-12-06, 18:03   #14
ET_
Banned

"Luigi"
Aug 2002
Team Italia

2·2,383 Posts

Quote:
 Originally Posted by feromant mmff v0.28 handles divisors of size up to 252 bits. For the given range will be used by the kernels mfaktc_barrett224_F160_191gs and mfaktc_barrett236_F160_191gs.
I should definitely take a break. I was referring to mmff 0.27, not mmff 0.28.

With mmff 0.28 the range should take 56 days at 100M k per second using a GTX580 or GTX980.

Luigi

2016-12-06, 18:06   #15
ET_
Banned

"Luigi"
Aug 2002
Team Italia

2×2,383 Posts

Quote:
 Originally Posted by rogue For these is there a limit for k where ppsieve is slower than fermfact?
AFAIK, ppsieve is always faster than FermFact, but you need FermFact to (at least) start a sieve, because of ppsieve's limitations on small p (iirc, p > kmax).

In other wrds, I usually start seving for 12/24 hours with FermFact, then save the remaining candidates and use the file as an input for ppsieve.

Last fiddled with by ET_ on 2016-12-06 at 18:08

2016-12-06, 19:25   #16
rogue

"Mark"
Apr 2003
Between here and the

22·1,481 Posts

Quote:
 Originally Posted by ET_ Code: 180000 15000 30000
I'll do this range. There are about 350 tests for it.

2016-12-06, 20:55   #17
feromant

"Roman"
Dec 2016
Everywhere

2·13 Posts

Quote:
 Originally Posted by ET_ With mmff 0.28 the range should take 56 days at 100M k per second using a GTX580 or GTX980
Luigi pay attention to ETA (1d16h = 40h). 40h*8exp = 320h/24 = 13 days. Given the increase in ETA with increasing N and get 15 days but not 56 days.
Attached Thumbnails

2016-12-06, 21:36   #18
rogue

"Mark"
Apr 2003
Between here and the

172416 Posts

Quote:
 Originally Posted by ET_ Code: 180000 15000 30000 no data available 180001-181742 13000 30000 no data available 181744-189999 13000 30000 no data available
I completed 180000 15000 30000. No primes in that range. For that n such a range of 15000 takes about 1 hour, but I was running on 4 cores on a fast laptop.

On my laptop I estimate about 96 days for 18001-181742 on a single core (after seiving) and about 487 days for 181744-189999 on my laptop.

Last fiddled with by rogue on 2016-12-06 at 21:36

2016-12-07, 11:47   #19
ET_
Banned

"Luigi"
Aug 2002
Team Italia

2×2,383 Posts

Quote:
 Originally Posted by feromant Luigi pay attention to ETA (1d16h = 40h). 40h*8exp = 320h/24 = 13 days. Given the increase in ETA with increasing N and get 15 days but not 56 days.
I had more conservative speeds for mmff, to cope with people having slower boards.

 2016-12-07, 11:51 #20 ET_ Banned     "Luigi" Aug 2002 Team Italia 476610 Posts OK, so here we have the actual reservation ranges: Code: 172-179 281000000000000 350000000000000 15 days with mmff, 35 days with Feromant_CUDA 6151-6199 400000000 2500000000 1500 days with a single-core pmfs 10000 253000000 269000000 19 hours with a single-core pmfs 10001 50000000 269000000 11 days with a single-core pmfs 10003-10999 50000000 269000000 5913 days with ppsieve_cuda & pfgw 14501-14999 30000000 70000000 640 days with ppsieve_cuda & pfgw 80001-80999 80000 100000 150 days with ppsieve_cuda & pfgw 114000-119999 40000 99999 12 years with ppsieve_cuda and pfgw 120001-129999 30000 40000 4 years with ppsieve_cuda & pfgw 130001-139999 20000 30000 5 years with ppsieve_cuda & pfgw 160001-169999 15000 20000 3 years with ppsieve_cuda & pfgw 180001-181742 13000 30000 96 days with ppsieve_cuda & pfgw 181744-189999 13000 30000 487 days with ppsieve_cuda & pfgw Thank you Mark for your help, I'm going to update the most wanted page. Roman, are you going to take N=172-179, or is it free for reservation? Luigi Last fiddled with by ET_ on 2016-12-07 at 11:52
2016-12-07, 15:48   #21
feromant

"Roman"
Dec 2016
Everywhere

328 Posts

Quote:
 Originally Posted by ET_ Roman, are you going to take N=172-179, or is it free for reservation?
Unfortunately, my GTX580 is now busy with other work and hardly will be released this year

 2016-12-07, 18:57 #22 rogue     "Mark" Apr 2003 Between here and the 10111001001002 Posts I'll take this range: 10000 253000000 269000000

 Similar Threads Thread Thread Starter Forum Replies Last Post ET_ Operazione Doppi Mersennes 489 2020-09-19 14:24 kar_bon Riesel Prime Data Collecting (k*2^n-1) 129 2016-09-05 09:23 R.D. Silverman NFS@Home 15 2015-11-29 23:18 R.D. Silverman Cunningham Tables 15 2011-03-04 21:01 paulunderwood 3*2^n-1 Search 15 2008-06-08 03:29

All times are UTC. The time now is 23:55.

Wed Sep 30 23:55:40 UTC 2020 up 20 days, 21:06, 2 users, load averages: 1.50, 1.34, 1.41