On Fri, Jul 19, 2019, 8:00 AM Ken Pettit wrote:
> On 7/18/19 6:20 PM, John R. Hogerhuis wrote:
>
> > I guess the value of (packed) BCD is multiply / divide by 10 is a bit
> > shift operation. And base 10 rounding operations are straightforward.
> >
>
> You don't even need a bit shift ... just
On 7/18/19 6:20 PM, John R. Hogerhuis wrote:
I guess the value of (packed) BCD is multiply / divide by 10 is a bit
shift operation. And base 10 rounding operations are straightforward.
You don't even need a bit shift ... just subtract 1 from the exponent
value. Also converting to ASCII
In article
,
John R. Hogerhuis wrote:
> I noticed Woz's 6502 code is well commented and on the web. It looks like
> it uses a binary mantissa.
If you're looking for an 8080 fp lib, the one extracted from Lawrence
Livermore Labs BASIC (LLLBASIC) has a pretty good rep. The one inside
Processor
I noticed Woz's 6502 code is well commented and on the web. It looks like
it uses a binary mantissa.
I guess the value of (packed) BCD is multiply / divide by 10 is a bit shift
operation. And base 10 rounding operations are straightforward.
-- John.
In article
,
John R. Hogerhuis wrote:
> Actually thinking some more about it, why would BCD floating point code
> be more compact even in ROM?
I haven't really looked, but the 8085 might be at nearly as efficient at
BCD as pure binary.
Also, the M100 was the "Micro Executive WorkStation",
I've wondered about that too - "Radix-100" FP used on TI CC40/TI74/TI95
programmables used an 8-byte format, byte 0 being the exponent biased
so that 40 = 0, and a 7-byte BCD mantissa.
...
On 7/18/19, John R. Hogerhuis wrote:
> Actually thinking some more about it, why would BCD
Actually thinking some more about it, why would BCD floating point code be
more compact even in ROM?
Seems like BCD floats would be inefficient in every way.
-- John.
In article <5d2f2bc9.4010...@gmail.com>,
Ken Pettit wrote:
> Hey John,
> I was reading the posts about posits ... pretty interesting.
> But the Model T ROM doesn't actually use IEEE floating point format ...
> It uses a format where the first byte contains the sign bit and 7-bit
>
Ah, all good points! I didn't figure the ROM used IEEE floating point, but
I didn't realize how different it was!
-- John.
Hey John,
I was reading the posts about posits ... pretty interesting.
But the Model T ROM doesn't actually use IEEE floating point format ...
It uses a format where the first byte contains the sign bit and 7-bit
exponent, followed by 3 or 7 bytes of BCD encoded data. For the
exponent, 40h
Nice to see there's still work being done on the low level parts of
compander science.
On Wed, Jul 10, 2019, 16:59 John R. Hogerhuis wrote:
> Something new under the sun?
>
> Tangentially related, but I thought this was interesting. The Model T ROM
> uses 32-bit single precision and 64-bit
11 matches
Mail list logo