C floating point number notation

Question

I noticed that this code compiles but I have no idea why:

int main() {
  double z = 0.000000000000001E-383DD;
}

But I'm not sure what the DD at the end of the number means. I've looked through the standard but there's no mention of this.

I got this number from the following command:

$ gcc -dM -E - < /dev/null
#define __DBL_MIN_EXP__ (-1021)
#define __FLT_MIN__ 1.17549435e-38F
#define __DEC64_DEN__ 0.000000000000001E-383DD
...

Might this be a GCC extension?

Answer 1

Right it's a GCC extension to indicate 64-bit decimal floating point literals.

Other extension suffixes:

• Complex literals, e.g. 1.0i, 1.0j, 1.0fi, etc. This purely a GCC extension and the standard C99 way is to use the macro I (1.0*I, 1.0f*I, etc.)

• Additional floating point types, which are purely a GCC extension:

  • 1.0w → __float80 (80-bit binary floating point)
  • 1.0q → __float128 (128-bit binary floating point)

• Decimal floating point types, which is based on the proposal N1312: Extension for the programming language C to support decimal floating-point arithmetic:

  • 1.0df → _Decimal32 (32-bit decimal floating point)
  • 1.0dd → _Decimal64 (64-bit decimal floating point)
  • 1.0dl → _Decimal128 (128-bit decimal floating point)

• Fixed-point types, which is based on the proposal N1169: Extensions to support embedded processors:

  • 0.5hr, 0.5r, 0.5ulr, etc. → _Fract types (fixed point types with magnitude ≤ 1)
  • 5.0hk, 5.0k, 5.0ulk, etc. → _Accum types (fixed point types)

Answer 2

Maybe it stands for a densely packed Decimal of 64 bits?

This wikipedia article section looks like it matches the name of the define, no?

Answer 3

Yes, it's a GCC extension to support decimal float. The literal has type _Decimal64, but is converted when assigned to the double variable.