double to sc_fix(ed) conversion

[scsc]

I dug up this old thread on double(float) conversion to sc_bv: 

 

Apparently we can convert a double to sc_ieee_double (see the snippet below from that post) then manually convert to sc_fix(ed) from mantissa and exponent or convert to sc_fixed from the bitvector (with some truncation because the bitvector would have different bitwidth to sc_fixed. But just wonder if there are built in methods that perform the conversion to sc_fixed directly? 

double d = ...;

sc_dt::scfx_ieee_double id(d); // convert to IEEE 754 bitfield

// prepare parts
bool               sgn = id.negative();
sc_dt::sc_uint<11> exp = id.exponent();
sc_dt::sc_uint<53> mnt = ( sc_dt::uint64( id.mantissa1() ) << 20 ) | id.mantissa0();

// concatenate parts to bitvector
sc_dt::sc_bv<64> bv = ( sgn, exp, mnt );

 

[Eyck]

The example you are quoting is conversion to bit vector representation. LRM sections 7.10.14 and 7.10.18 specify sc_fix(ed) and there you find constructors accepting double as well as assignment operators for double. So you can do a direct assignment.

[scsc]

Thanks Eyck,

I tried to use double input for an sc_fixed data type (see the attached example). Somehow the direct assignment didn't work. Or because the errors I got were from sc_port.h, which has not been used at all. So I wasn't sure the direct assignment actually works.

 

#define SC_INCLUDE_FX
#define N 25
#include <systemc.h>

using namespace sc_core;
using namespace sc_dt;
typedef sc_fix<32, 16, SC_RND> fixed_type;

double fix_fir(double _in[])
{
    
    fixed_type in[N], c[N], t[N], y;

    int i;
    double ct = 0.9987966;
    for (i = 0; i < N; i++) {
        in[i] = _in[i];
        c[i] = ct;
        ct = ct / 2;
    }

    for (i = 0; i < N; i++) {
        t[i] = c[i] * in[i];
        y += t[i];
    }
    return y;
}

int sc_main(int argc, char* argv[]) {

    double _in[25];
    double y;
    for (int i = 0; i < 25; i++) {
        _in[i] = (double)i;
    }

    y = fix_fir(_in);
    cout << "y = " << y << endl;

    return 0;
}

 

Some errors are like these:
Error (active)    E0706    expected a ',' or '>'    sc_fixed_fir   \Documents\systemc-2.3.2\src\sysc\communication\sc_port.h  386
Error (active)    E0306    default argument not at end of parameter list    sc_fixed_fir    \Documents\systemc-2.3.2\src\sysc\datatypes\fx\sc_fixed.h    61 

Line 61 of sc_fixed.h has this definition in which the default argument to sc_fixed is "int N":

// classes defined in this module
template <int W, int I, sc_q_mode Q, sc_o_mode O, int N> class sc_fixed;

[Eyck]

Your problem is the

#define N 25

before including systemc.h since N is used as template parameter name in some systemc class templates.

I put your code on edaplayground: https://www.edaplayground.com/x/s6Fu with a few minor changes and there it compiles and works....

[scsc]

9 hours ago, Eyck said:

Your problem is the

#define N 25

before including systemc.h since N is used as template parameter name in some systemc class templates.

I put your code on edaplayground: https://www.edaplayground.com/x/s6Fu with a few minor changes and there it compiles and works....

Thanks again Eyck. I confirmed that #define N was the culprit. I can run the example after changing N to something else (along with a typo in the typedef line). This is a tricky one. But the errors made sense looking back as there is "int N" in sc_fixed template

[Eyck]

That's why it always better to use an unnamed enum as shown in the edaplayground example. This is typesafe and allows the compiler to check things.