Le 12/08/2013 12:19, Julien Puydt a écrit : > Le 10/08/2013 22:16, Volker Braun a écrit : >> Since the factorization starts with 2^64 it looks like there is >> something wrong with how uint64_t is being used in the computation. >> Possibly a 32-bit/64-bit problem. > > The following computation is even more interesting: > > sage: n=2^100-1;n > 1267650600228229401496703205375 > sage: pari(n) > 31668760015517046052350966685377757183 > sage: pari(n) > 27806445565810435977365220706566012927 > sage: pari(n) > 28935765463492261505845880393861955583 > sage: pari(n) > 67017549253314407993604807848961769471 > > I know the build is quite experimental, but this is off-limit: there's > something extremely fishy being done! > > I see two possibilities: either something is using the clock to convert > integers to integers, or some piece of code has been making evil things > with the stack, and has been getting away with it until now by sheer luck! > > The second is my suspect...
If it had been a moving stack, I guess running the following several times would have given a segfault: for _ in range(1000): for _ in range(1000): pari(n) since it didn't... I'm back to trying to understand what can happen... Snark -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.