20110907, 19:46  #45  
"Robert Gerbicz"
Oct 2005
Hungary
11^{2}·13 Posts 
Quote:
When I put a code, that means it survived thousands of different runs for different st,en,interval inputs for en<10^5, and passed a large test, where en~1e10, just to check there is no int/long long int problem in the code. Last fiddled with by R. Gerbicz on 20110907 at 19:51 

20110907, 19:58  #46  
Jun 2003
Ottawa, Canada
3×17×23 Posts 
Quote:
I have 8GB of RAM on that machine, but I never saw it use more than 1.4GB of RAM though. I'm now retrying the range with the wilsontest2 code using MPIR. There also might have been a bad GMP5 library or something since my old mingw64 environment was misbehaving. Jeff. 

20110907, 20:19  #47  
"Robert Gerbicz"
Oct 2005
Hungary
3045_{8} Posts 
Quote:
Yes, that uses less than 5 GB, the reason that your 71e970e9=1e9 range contains much less than 2*interval primes. Last fiddled with by R. Gerbicz on 20110907 at 20:21 

20110908, 21:05  #48 
Mar 2011
Germany
97 Posts 

20110909, 00:00  #49 
Jun 2003
Ottawa, Canada
3×17×23 Posts 
Some results with different intervals:
interval=50000128 Done the st=71000000000en=72000000000 interval. Time=57585 sec. interval=5000192 Done the st=72000000000en=73000000000 interval. Time=111822 sec. Huge difference in speed with those intervals. Testing 75000064 right now. 
20110909, 05:23  #50  
Mar 2011
Germany
141_{8} Posts 
Quote:
Reason is that there are only 40M primes in that interval. So except from other effects the speed should be the same. 

20110909, 12:38  #51 
(loop (#_fork))
Feb 2006
Cambridge, England
14466_{8} Posts 
It will be faster; the interval parameter is used in two places in the software, in one place min(interval length, number of primes) is relevant, and in the other it isn't.
Compare the timings in my posts http://www.mersenneforum.org/showpos...6&postcount=12 and http://www.mersenneforum.org/showpos...5&postcount=35 in both of which the interval is larger than the number of primes in the range. To compare: Code:
interval RAM time slices 1e8 3.647 79427.6 3 2e7 1.904 97210.0 3 1e7 0.983 126835.4 4 4e6 0.397 200194.4 11 Last fiddled with by fivemack on 20110909 at 14:07 
20110909, 13:15  #52  
"Bob Silverman"
Nov 2003
North of Boston
2^{3}·3·311 Posts 
Quote:
very very rare. Much rarer than than Mersenne primes. They should be comparable in density (heuristically) to Wiefrich or Miramanoff primes. As John Selfridge said: We know that loglog N goes to infinity. However, noone has ever actually observed it doing so. 

20110909, 13:32  #53  
"Tapio Rajala"
Feb 2010
Finland
3^{2}×5×7 Posts 
Quote:
Last fiddled with by rajula on 20110909 at 13:47 Reason: corrected some of the (possibly many) typos, limits, numbers, errors... 

20110909, 13:43  #54  
"Mark"
Apr 2003
Between here and the
3·17·131 Posts 
Quote:


20110909, 13:53  #55  
"Bob Silverman"
Nov 2003
North of Boston
1110100101000_{2} Posts 
Quote:
represents a faction with respect to some other value. The expected number (which is an absolute number) of Wilson primes up to N is ~C loglog(N), giving .08C as the number of expected primes in [1e12, 1e13] for some constant C. I have never seen a strong argument that gives the value of C. Is it 1?? 

Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Twin prime search?  MooooMoo  Twin Prime Search  115  20100829 17:38 
k=51 or about coordinated prime search  Kosmaj  Riesel Prime Search  7  20070713 22:15 
Prime Search on PS3?  Kosmaj  Riesel Prime Search  6  20061121 15:19 
Genetics and Wilson's theorem  David John Hill Jr  Science & Technology  2  20060510 14:10 
Generalized wilson's theorem  bouayoun  Math  3  20040312 18:24 