Error : Constant Operands

Hi

Iam implementing following code as show below

m = 10;
v = [1 1 2 : m]

v = 1×11

1 1 2 3 4 5 6 7 8 9 10

If I try to make fixed point , I am getting the error "Operands must be constants." I don't understand what is expected here. Can someone help me ?

Thank you

[Geoff Hayes]

@Jogger - please clarify what you mean by "try to make this fixed point". What line of code leads to the error message?

[Life is Wonderful]

try to make this fixed point - Sorry I didn't write my requirment clearly.

I mean m is a Embedded.fi object here . Let say m = 10 , then class of m is 'double' but when I use numerictype or fi, then m is translated to 'embedded.fi'.

The problem i see colon operator doesn't seems to work for 'embedded.fi'.

Answer 1

[David Hill]

Does changing to double work? I don't have the toolbox

v = [1 1 2 : double(m)]

[Life is Wonderful]

Thanks a lot, it works

Answer 2

[Andy Bartlett]

I'm puzzled by the error you described "Operands must be constants.".

When I attempt to reproduce the issue, I don't observe any problem.

I created this simple function.

function y = makeVec(m)
    y = [1 1 2 : m];
end

When I test this with double or fixed-point inputs

u1 = 10;
u2 = fi(10);
y1 = makeVec(u1)
y2 = makeVec(u2)

the outputs are as expected.

y1 =
     1     1     2     3     4     5     6     7     8     9    10
y2 = 
     1     1     2     3     4     5     6     7     8     9    10

          DataTypeMode: Fixed-point: binary point scaling
            Signedness: Signed
            WordLength: 16
        FractionLength: 0

Code generation can provide additional restrictions, so I tried that too.

codegen makeVec -args {u1} -o makeVec1_mex
codegen makeVec -args {u2} -o makeVec2_mex

y1m = makeVec1_mex(u1)
y2m = makeVec2_mex(u2)

Again, the outputs were as expected.

y1m =
     1     1     2     3     4     5     6     7     8     9    10
y2m = 
     1     1     2     3     4     5     6     7     8     9    10

          DataTypeMode: Fixed-point: binary point scaling
            Signedness: Signed
            WordLength: 16
        FractionLength: 0

If you could clarify the original reproduction steps, that would be appreciated.

Andy

[Life is Wonderful]

Thanks @Andy Bartlett,

The problem i see is coming from factorial function in matlab . Factorial function is unsupport by the fixed point designer . So I wrote my own implementation. I reliased the mistake in declaring fi Object and double togeather.