In my good old days (1963) I used FORTRAN II on an IBM 1620 but after a
few months it was replaced by an Elliott 503 machine that used 39-bit
words, and stored one floating-point number in each.
Its floating-point form was x = a*2**b such that -1 <= a <= 1/2 or a = 0
or 1/2 <= a < 1, and -25 < b < 256. We used Algol on it; there was said
to be FEAT, a Fortran-to-Elliott Algol Translator, but I never got it
to work: it was easier to translate my few programs manually to Algol.
I don't know if double precision was available on that machine.
On Thu, 6 Apr 2017, Keith Bierman wrote:
> Date: Thu, 6 Apr 2017 18:56:39 -0600
> From: Keith Bierman <[log in to unmask]>
> Reply-To: Fortran 90 List <[log in to unmask]>
> To: [log in to unmask]
> Subject: Re: COMP-FORTRAN-90 Digest - 5 Apr 2017 to 6 Apr 2017 (#2017-7)
>
> On Thu, Apr 6, 2017 at 5:00 PM, COMP-FORTRAN-90 automatic digest system <
> [log in to unmask]> wrote:
>
>> In the good old days before those good old days, DOUBLE PRECISION was 64
>> bits.
>
>
> Actually, in the really good old days, it was 72 (Univac 2x36). Sometimes
> it was 60 (that might have been a site specific adjustment to the CDC,
> which logic says ought to have been 120 and I'm sure it was at other
> sites). 64 was mostly true on the most numerically deplorable (IBM Hex
> representation).
>
>
>
>
> Keith Bierman
> [log in to unmask]
> 303 997 2749
>
-- John Harper, School of Mathematics and Statistics
Victoria University, PO Box 600, Wellington 6140, New Zealand
e-mail [log in to unmask] phone (+64)(4)463 5276 fax (+64)(4)463 5045
|