Re: Looking For Simple Routine To Put Integer Into Float

Re: Looking For Simple Routine To Put Integer Into Float

[Ludovic Brenta]

ChristopherL <clusardi2k@aol.com> writes:

> Can you modify the below code to convert an integer variable of 32
> bits to a variable of type float.

Why does it matter that the integer variable is 32 bits?

How many bits does the float type use?

> type Unknown_integer_type is range -(2**16) ..(2**16) -1; -- 32 Bit
> Integer

This looks like a 17-bit integer type to me, not 32-bit.

> An_Unknown_Integer: Unknown_integer_type;
> An_Integer: Integer;
> A_Float: Float;            -- Hoping this float is usually 32 bits

Then it cannot hold a 32-bit integer without losing some precision,
since some of the float's 32 bits will be reserved for the exponent.
But since the integer type is only 17-bit wide, that's OK; the mantissa
of a 32-bit float is very probably wider than 17 bits.

> An_Integer := 1234;
>
> An_Unknown_Integer := An_Integer;
>
> A_Float := Float (An_Unknown_Integer);

That's a type conversion, it assigns the value 1234.0 to A_Float.  Since
you ask a question, I can only assume that that's now what you wanted.
So what is it that you want, exactly?

-- 
Ludovic Brenta.

Re: Looking For Simple Routine To Put Integer Into Float

[Jeffrey R. Carter]

ChristopherL wrote:
> 
> type Unknown_integer_type is range -(2**16) ..(2**16) -1; -- 32 Bit
> Integer

If you want an unconstrained, 32-bit, signed integer type, something from 
Interfaces (such as Integer_32) is usually the way to go.

> An_Unknown_Integer: Unknown_integer_type;
> An_Integer: Integer;
> A_Float: Float;            -- Hoping this float is usually 32 bits

Why hope? If you need a 32-bit floating-point type, again something from 
Interfaces might be in order, or declare an appropriate type for your application:

type Real is digits 6 [range Low .. High];
for Real'Size use 32;

In any case, the size of Float should be documented for your compiler.

> An_Integer := 1234;
> 
> An_Unknown_Integer := An_Integer;

This shouldn't compile, as An_Integer has a different type than 
An_Unknown_Integer; you need a type conversion.

> A_Float := Float (An_Unknown_Integer);

What is the point of Unknown_Integer_Type and An_Unknown_Integer? What's wrong with

A_Float := Float (An_Integer);

?

-- 
Jeff Carter
"You can never forget too much about C++."
115

Re: Looking For Simple Routine To Put Integer Into Float

[Stephen Leake]

ChristopherL <clusardi2k@aol.com> writes:

> Can you modify the below code to convert an integer variable of 32
> bits to a variable of type float.
>
> 
>
> A_Float := Float (An_Unknown_Integer);

This answers your question as asked; it converts any integer type to the
Float type.

Perhaps you want to ask something else?

-- 
-- Stephe

Re: Looking For Simple Routine To Put Integer Into Float

[mockturtle]

On May 25, 4:47 am, Stephen Leake <stephen_le...@stephe-leake.org>
wrote:
> ChristopherL <clusard...@aol.com> writes:
> > Can you modify the below code to convert an integer variable of 32
> > bits to a variable of type float.
>
> > A_Float := Float (An_Unknown_Integer);
>
> This answers your question as asked; it converts any integer type to the
> Float type.
>
> Perhaps you want to ask something else?

My theory: maybe the OP wanted to do a "bitwise copy" of the 4 octets
of the Integer into the 4 octets of the Float? Something like an
Unchecked_Conversion?  (This would explain the "Hoping this float is
usually 32 bits" remark)
>
> --
> -- Stephe