I don't get why in the attached code, if line 36 I set transa to 't', then I get a segmentation fault. The matrix A is symmetric, so there shouldn't be any other changes to do to make this work ? In its Fortran version, dcsr_multiplication.f, there is no problem when I switch trans line 82.
Thanks in advance for any help.
qweasd q. wrote:Thanks for all your help !
Another quick question, is it planed to support dcsrmultcsr with arbitrary 'matdescr' (for example symmetric input, unsymmetric output). I'm using this routine to compute A*D where A is symmetric (and so only its lower part is stored) and D is diagonal (but might as well be a general matrix). What I'm doing right now is csradd(csrmultcsr(A, D), csrmultcsr(A', D)) (I don't forget to remove the diagonal terms of csrmultcsr(A', D)), but not only it is making the MKL segfaults (not anymore with your fix though), I don't think this is very efficient.
It seems Update 3 breaks ARPACK. Because I also use it in my application, I won't update until Update 4. Thanks anyway, I hope everything will be alright in the next update.