error output numbers

error output numbers

[w]

when i use ada to write a program to calculate factorial, i found that
result after 12! is always wrong.

for example, 

12! = 479001600
but 13! gives me 1932053504.    
The correct result should be 6227020800.

May be there are some problems with my integer declaration?
I decleared all outputs to be INTEGER.

Thanks!       :-)

Re: error output numbers

[Marin David Condic]

The correct result is in excess of six billion. A garden variety integer is
only good to 2**31 - or around two billion - on most Ada implementations.
Hence your calculation is overflowing. You did not say what compiler you are
using but I bet if you check you'll discover there is a switch you can set
to turn on runtime checks for overflows. Do that, and your program will
probably raise Constraint_Error.

If your compiler supports a longer integer, this might help your
calculations, but it is only a matter of time before this will overflow too.
Factorials grow very rapidly to exceed whatever size integer you want to
pick.

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com
Web:      http://www.mcondic.com/



"w" <wwminirl@hotmail.com> wrote in message
news:01c0b0b5$af29df60$2a9ec594@default...
> when i use ada to write a program to calculate factorial, i found that
> result after 12! is always wrong.
>
> for example,
>
> 12! = 479001600
> but 13! gives me 1932053504.
> The correct result should be 6227020800.
>
> May be there are some problems with my integer declaration?
> I decleared all outputs to be INTEGER.
>
> Thanks!       :-)

Re: error output numbers

[tmoran]

> > I decleared all outputs to be INTEGER.
> The correct result is in excess of six billion. A garden variety integer is
> only good to 2**31 - or around two billion - on most Ada implementations.
  This seems an excellent demonstration of the advantages of using your own
declared integer types, rather than whatever your current compiler happens
to be offering today.  If the declaration was
  type Factorial_Values is range 1 .. 999_999_999;
it would be clear that a value of a billion or more won't work.  OTOH,
  type Factorial_Values is range 1 .. 999_999_999_999;
would let the compiler know it better use more than 32 bits, and it
would be clear exactly how big a factorial value would still work.

Re: error output numbers

[David C. Hoos, Sr.]

13! is greater than 2**31 -1, so type Integer cannot be used to
hold such a value.

You can use a 64 bit modular type (mod 2 ** 64) to extend the range
somewhat.

"w" <wwminirl@hotmail.com> wrote in message
news:01c0b0b5$af29df60$2a9ec594@default...
> when i use ada to write a program to calculate factorial, i found that
> result after 12! is always wrong.
>
> for example,
>
> 12! = 479001600
> but 13! gives me 1932053504.
> The correct result should be 6227020800.
>
> May be there are some problems with my integer declaration?
> I decleared all outputs to be INTEGER.
>
> Thanks!       :-)

Re: error output numbers

[Chad R. Meiners]

Perhaps you should compile with the -gnato option ;)
(assuming of course you are using GNAT)

-Chad R. Meiners

"w" <wwminirl@hotmail.com> wrote in message
news:01c0b0b5$af29df60$2a9ec594@default...
> when i use ada to write a program to calculate factorial, i found that
> result after 12! is always wrong.
>
> for example,
>
> 12! = 479001600
> but 13! gives me 1932053504.
> The correct result should be 6227020800.
>
> May be there are some problems with my integer declaration?
> I decleared all outputs to be INTEGER.
>
> Thanks!       :-)

Re: error output numbers

[Nils]

This is 32 bit integer problem, note that
2**32 = 4294967296
+ 1932053504.=
6227020800

try interfaces.integer_64 if you are using gnat
or define your own type

nils

w wrote:

> when i use ada to write a program to calculate factorial, i found that
> result after 12! is always wrong.
>
> for example,
>
> 12! = 479001600
> but 13! gives me 1932053504.
> The correct result should be 6227020800.
>
> May be there are some problems with my integer declaration?
> I decleared all outputs to be INTEGER.
>
> Thanks!       :-)