3 or 6
bytes ?
-----Message d'origine-----In a message dated 11/03/03 03:40:54 GMT Standard Time, [EMAIL PROTECTED] writes:
De : [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Envoyé : mercredi 12 mars 2003 16:00
À : [EMAIL PROTECTED]
Objet : Re: [ql-users] Sbasic and numbers
As a final digression, I've always bemoaned the fact that floating point
is only allowed a twelve bit exponent. It could have gone for 15 bits,
which would have reduced the overall code and increased the speed!
There you go.
Ah. But you could use the FPU, with its 15-bit exponent and 64-bit mantissa, on the Q40, Q60 and some QXLs to do rather longer precision arithmetic. Actually the FPU on a Thor 21 was used even to do Qdos arithmetic although AFAIK only Simon Goodwin's trig functions (on the Q40 and Q60) actually use the FPU. Otherwise Qdos is just the old fashioned 3-byte 32 bit mantissa, 12 bit exponent.
George
