I've fixed the known issues and have released 3.5.2. The linux build isn't available quit yet, but should be posted later today. Sorry Dave, you'll have to...
... Hello Mark, Even with this fix, the crash of pfgw64 (windows version) I reported to you last friday is still there (visual studio debugger still detects...
... I presume he didn't build with the correct code: pfgw64 -V -q"nextprime(2^62)" PFGW Version 3.5.2.64BIT.20110926.Win_Dev [GWNUM 26.6] Switching to...
... as requested, though I'm afraid it's "more" rather than "better" up to 50 digits: [1.988, 2.056, 1000] [1.916, 2.028, 1000] [1.848, 1.959, 1000] [1.923,...
... That timing is awful: 148 seconds for what Pari-GP can do 1000 times in 70 milliseconds: for(k=1,1000,isprime(nextprime(2^62))); print(gettime" ms"); 72 ms...
... I had sent similarly inconclusive results, up to 300 digits, to David Baugh, off list. With only 5000 attempts, at each size, it would be a waste of space...
pfgw64s -V -q"nextprime(2^63-2^41)" PFGW Version 3.5.2.64BIT.20110926.x86_Dev [GWNUM 26.6] Sieve re-allocated with a limit of 18446739675663040512. Signal...
... I agree, but I have a question. First, does Pari-GP guarantee that the nextprime function returns a prime or could it return PRPs? I use the new sieving...
... ?nextprime nextprime(x): smallest pseudoprime >= x. And that means a BPSW PRP, so you will become very famous if you detect a failure of GP#s nextprime(), ...
... I'll modify pfgw to use different code behind the nextprime function then. As for the other bug with -q"nextprime(2^63-2^41)", I have fixed it (but not...
... I've checked this on both MSVS 2005 and 2010 and both have the issue. However, on an old Linux RedHat 3.0 with a gcc 3.2.3, the 32 bit executable output...
... Only if he needs primes > 2^62. Does anyone see a need to fix the double to uint64 bug in Win32? If I use different code for nextprime, I don't think that...
... Version 3.5.2 is sick for much smaller primes than that. Just look at this ABC2 file: $ more tmp ABC2 nextprime(p($a)) a: from 999990 to 1000000 $ pfgw352...
... That is only in the Linux build. I had fixed it, but Steven built the release without the current code. Steven, please get the latest code (this fix is...
with quite dense granularity of 1e9 are available there: http://www.primefan.ru/stuff/primes/table.html I'm going to reach 1e17, so the archive will be updated...
... Well, on Windows, I still have problem with the lastest 3.5.2 with primes much smaller that 2^62. The following script crashes pfgw64 as soon as I change...
... Does that include the bug fix I suggested to you yesterday? It does not crash for me with that fix. If this is a show-stopper for you, I can post a...
... These entries are correct. The next one should be about 1.882724e20; I have to hack Xavier Gourdon's program to find the exact value :-) Best regards, ...
... You're right. This was without what you suggested to me yesterday. I have just followed your suggestion and added a couple of m_IndexInWindow = false; in...
... It's slow because it uses the sieve code to get the next prime. The way you have it in that script it would be faster to use the p() function. I'm working...
... 8 years ago, Pari_GP's ispseudoprime and nextprime were easy to fool: http://tech.groups.yahoo.com/group/primenumbers/message/11027 but now they use a...
... If someone wants to code a BPSW test for pfgw, I would welcome their effort. I suspect someone on this list (Phil?) might have already written one that...
