A convenient notation for floating point and integer literals with a
specific precision, e.g. 4_f16 or 1_i32. This is particular useful
when programming accelerators since performance there is often much
higher when using fewer bits.
These suffixes are supported:
julia> using FloatIntLiterals
julia> 1_i8
1
julia> 1_i16
1
julia> 1_i32
1
julia> 1_i64
1
julia> 1_i128
1
julia> 1_u8
0x01
julia> 1_u16
0x0001
julia> 1_u32
0x00000001
julia> 1_u64
0x0000000000000001
julia> 1_u128
0x00000000000000000000000000000001
julia> 1_f16
Float16(1.0)
julia> 1_f32
1.0f0
julia> 1_f64
1.0
julia> 1.1_f16
Float16(1.1)
julia> 1.1_f32
1.1f0
julia> 1.1_f64
1.1
julia> 1.1e-4_f16
Float16(0.00011)
julia> 1.1e-4_f32
0.00011f0
julia> 1.1e-4_f64
0.00011