[VA] SRC #015b - HP-15C & clones: COMPLEX Matrix Inverse up to 8x8

[Werner]

Hi Gerson,
To solve a 4x4 system on an original 15C, you have to leave off the inversion code LBL E. I see you used up 9 registers for programs, so you used the full 63 bytes and then there’s not enough room left to solve the system. LBL C and D together only take 48 bytes so less than 7 registers, and then it will work (with 1 DIM (i) ).
And yes, when A+B is singular you get a meaningless answer, unfortunately. The example you gave however, is rank-deficient anyway as B=0 and det(A)=0, so in this case there is no workaround, save a Least Squares solution, which is a different beast altogether.
Cheers, Werner