matrix - Eigen Library : real time processing -

matrix - Eigen Library : real time processing -

i'm trying utilize eigen no heap allocation real time processing.

as far know:

matrix<double, dynamic, dynamic> allocates on stack matrix<double, dynamic, dynamic, autoalign, 1, 1> allocates on heap , matrix<double, 1, 1> allocates on stack;

one way utilize eigen in real time resize matrices at startup, big size, using fraction of size.

my question : functions : rm.colpivhouseholderqr().solve(rhs), how can perform such computation dynamic matrices? function take whole size of matrix operation not want portion of size treated. can't use .resize()

so how perform operations real-time processing?

real-time processing involves same work beingness done repeatedly, on fixed schedule. 1 approach dynamic memory allocation have contiguous chunk (a pool) of memory linearly allocate from. mark of memory free @ end of processing period. allocation cost o(1), deallocation cost o(1), , given amount of work can guarantee memory allocations satisfied. so, 1 solution utilize custom allocator.

matrix real-time eigen