Solver, Inverting, FFT ect...

Here's FFT source code:

DECLARE FUNCTION bitRev&(j&,nu&)

FUNCTION bitRev&(j&,nu&)
    j1&=j&
    k&=0
    FOR i&=1 TO nu&
        j2&=j1&\2
        k&=k&*2+j1&-2*j2&
        j1&=j2&
    NEXT i&
    FUNCTION=k&
END FUNCTION

SUB newFFT(ftr AS DOUBLE PTR, fti AS DOUBLE PTR, jmax&,nu&)
    n2&=jmax&\2
    nu1&=nu&-1
    k&=0
    dw#=2*pi#/jmax&
    FOR l&=1 TO nu&
        pt&=1
        SHIFT LEFT pt&,nu1&
        DO WHILE k&<jmax&
            FOR i&=1 TO n2&
                p#=bitRev(k&\pt&,nu&)
                arg#=p#*dw#
                c#=COS(arg#)
                s#=SIN(arg#)
                trl#[email protected][k&+n2&]*c#[email protected][k&+n2&]*s#
                tim#[email protected][k&+n2&]*c#[email protected][k&+n2&]*s#
                @ftr[k&+n2&][email protected][k&]-trl#
                @fti[k&+n2&][email protected][k&]-tim#
                @ftr[k&][email protected][k&]+trl#
                @fti[k&][email protected][k&]+tim#
                INCR k&
            NEXT i&
            k&=k&+n2&
        LOOP
        k&=0
        nu1&=nu1&-1
        n2&=n2&\2
    NEXT l&
    k&=0
    DO WHILE k&<jmax&
        i&=bitRev&(k&,nu&)
        IF i&>k& THEN
            trl#[email protected][k&]
            tim#[email protected][k&]
            @ftr[k&][email protected][i&]
            @fti[k&][email protected][i&]
            @ftr[i&]=trl#
            @fti[i&]=tim#
        END IF
        INCR k&
    LOOP
END SUB

The code executes in realtime in PBCC10 (SB Input, double-buffered)
Converted from Pascal (original program / source PSK31 by Peter Martinez, I think)




-------------
Daniel
[email protected]

[E Dingsor]

Peter,
Go to http://www.ulib.org/webRoot/Books/Numerical_Recipes/

There You'll find Fortran or C++ code. It's easy to translate to Basic. They used to have Neumerical Recipes for Basic, but I'm not sure if it still in stock.
Anyway, Numerical REcipes is a need to have book if You do a lot of math programming.

Best regards

Eigil

Has anybody seen PowerBasic-algoritms for:

1) solving non-linear systems
2) solving (large) linear systems
3) FFT

Regards
Peter