c++ - Looking for one linear algebra library for embedded systems (without malloc and free) -
i use work microcontrollers. rtoss employ in applications not have free , malloc (and other calls assert), available, prefer have static in system.
i have started employ linear algebra, of them need dynamic memory. matrices dense , 'small' (not more 10x10).
i eigen (everything can static decided in compile time), apparently there bug calling asserts not provided rtos (even -dndebug). library should provide matrix decomposition routines (like qr, cholesky, lu...)
i prefer c instead c++. suggestions? many in advance!
anything wrong clapack? or straight fortran lapack (you can compile gfortran, part of gcc).
[c]lapack's routines take memory buffers in arguments allocated, , not heap allocation whatsoever. routines take "work" buffers in addition other arguments (for example, dgesdd computing svd), can call them special "size only" argument, , in response required size work buffers, can allocate wish.
Comments
Post a Comment