mabshoff wrote: > Hello Jaap, > > I assume you care about the following (computed on sage.math): > > sage: dance(11) > h^11 - 44*h^10 + 1045*h^9 - 16500*h^8 + 187935*h^7 - 1595748*h^6 + > 10199343*h^5 - 48691500*h^4 + 169140180*h^3 - 405230320*h^2 + > 600311624*h - 415232800 >
Whow! This is great! I can now updat the corresponding entry in the OEIS. My next wish is to find a formula for this kind of permanents! I'm improving speed of the methods in matrix2.pyx in the meantime. I'm makeing progress. >> sage: time dance(9) >> h^9 - 27*h^8 + 414*h^7 - 4158*h^6 + 29421*h^5 - 148743*h^4 + 530796*h^3 - >> 1276992*h^2 + 1866384*h - 1255608 >> CPU times: user 1453.75 s, sys: 8.03 s, total: 1461.78 s >> Wall time: 1487.04 Actually I'm running dance(10) now on my machine and I hope there will be no segfaults, because I simplified the code. It will take less than 3.5 hours to finish. > I am currently computing dance(12) on sage.math. William: feel free to > kill pid 5655 in case you consider this a waste of your resources. > > I had to kill the valgrind process on my local box because valgrind > stopped counting errors after the 1,000,000th. I also extrapolated how > long it would take and I am not willing to wait a month for it to > finish (if it does at all). > Cheers!!! > I started looking at the code itself to see if we cannot come up with > an easier to run test case that segfaults and currently I am looking > at the components like permanent() and so on that are using via > rook_vector. So if you have any ideas where that code could overflow > on 32 bit please let me know. > There are some issues related to the very art of permanents. There is no one who want to compute the permanent of a matrix larger than an int x int! In the literature you only will find calculations of certain types of (0,1) matrices larger than 50 x 50, except the trivial ones. I did send a patch for trac #931 to speed up the _choose part, and also boost things up with a local variant of the binomial function, more to come!(?) Jaap --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---
