[VA] SRC #014 - HP-15C & clones: Accurate NxN Determinants

[Werner]

.. and, my 15C version.
- works for n=1, too
- twice as fast
- uses 2n fewer matrix elements
- uses R0 and R1 only (not I)
- 45 bytes instead of 57

001 LBL B
002 RCL MATRIX A
003 STO MATRIX B
004 RESULT B

005 LBL 3
006 RCL DIM B           -- B is ixn
007 X<>Y
008 STO 0
009 STO 1
010 RCL B               -- recall Bii
011 DSE 0               -- i := i-1;
012 ISG 1               -- skip
013 RTN                 -- if B is single row, we're done
014 CHS                 -- sum := -Bii;
015 RCL 0
016 R^
017 DIM B               -- remove last row, which will be zeroes anyway
018 RCL MATRIX B
019 STO MATRIX E
020 R^                  -- put sum back into stack reg. X
021 GTO 0

022 LBL 2
023 0
024 LBL 1               -- zero out row i of E, j=i-1..1
025 STO E
026 DSE 1
027 GTO 1
028 X<>Y
029 DSE 0

030 LBL 0
031 RCL 0
032 STO 1
033 X<>Y
034 STO E               -- Eii := sum;
035 RCL- B              -- sum := sum - Bii;
036 DSE 1
037 GTO 2
038 RCL MATRIX E
039 RCL MATRIX A
040 x
041 CHS
042 GTO 3

Cheers, Werner