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

[Gerson W. Barbosa]

(10-03-2023 07:47 AM)Werner Wrote:  In the meantime I updated my post with a new, shorter and faster (for solving, at least) version.
It seems that every time I look at my code I find ways to improve it..

Thanks again!

Now we have a really pocket calculator suitable for this kind of task, even the original HP-15C is fast enough to handle these.

It took about 30 seconds to solve my second example (20.5 s +9.5 s). Basically the same results.

C =
| 15.7426645800 | [-1 ULP]
|-15.8956234800 | [+2 ULP]
| 0.07647945442 | [+8 ULP, previously 4 ULP]

D =
|-9.352382547 | [-5 ULP]
|-8.958908239 | [-7 ULP]
| 9.155645395 | [-4 ULP]

or

I₁ = 18.31115909∠-30.71364707° [-1, 0 ULP]

I₂ = 18.24644849∠-150.5940233° [-2, 0 ULP]

I₃ = 9.155964816∠89.52140479° [-3, 0 ULP, previously -6, 0 ULP]

ULP comparisons with the HP-GX results which are assumed to be more accurate, but not granted to be always exact to 12 digits.


Edited to fix formatting