Tag Info

New answers tagged

0

It would be easier to answer if you gave some examples, but I'll try anyway. As I understand your question, you allocate a large matrix in Fortran and set elements of it to 0 or 1 during execution. In Mathematica, you start with an empty list and append to it during exection. That would require lots of internal allocation and copying and doesn't reflect ...


4

This is a bit long for a comment. You can use a tandem of LinearSolve and NullSpace. But for exact problems this will, I'm fairly sure, use dense matrices. That takes you back to what Solve is doing anyway, under the hood. Assuming your inputs are integer or rationals you might avoid dense matrix algebra as follows. Use numerical methods to find a single ...


3

Ok, here you have it, but in the future this kind of questions will most likely be closed because the main problem arises from very basic errors. You can't expect others do the debugging for you. Take a look at the functions' definitions. {UT1, RA1, MM1, Z, MU, a1, L} = Transpose[RandomReal[{0, 1}, {10, 7}]]; dt = 1; rvsum1[1] = 5320007.301; rvsum1[i_] := ...


5

Two possible solutions are Use MySQL, SQLite or any other database that is not an in-memory database. HSQL keeps all its data in the memory at all times. I'm leaving this for future visitors since there is a version of HSQL that is built in that is in-memory, although that was not the case for the OP. Increase the Java heap size by using the same technique ...



Top 50 recent answers are included