On 4/13/2016 11:16 AM, Kustaa Nyholm wrote:
>
> On 13/04/2016 05:07, "KHMan" wrote:
>
>> bit MCU program that uses 16 bit floats, or
>> something very close in spirit...
>
> I posted this before:
>
> http://www.sparetimelabs.com/densitometer/densitometer.php
>
> That is a story of a densitometer (film optical density measurement
> device) that was implemented in an 8 bit (68HC11) assembler
> using 24 bit floating point, hand crafted in assembly.
>
>
> So I think this should qualify as something 'very close in spirit'.
>
> The source code is there too in case someone is interested.
>
> Even though I had to write the FP myself I thought at the
> time that using FB was the easies way to implement this.
>
> This device measures the amount of light passing through a
> strip of movie film to access the quality of the film development
> process, which is calculated at log10 of (I0/I) where I is
> the amount of light passing through the film under test
> and I0 is the same measurement without the film in the
> measurement chain.
>
> The range of I0/i is ~ 1-10000 for film densities of 0.01 to 4.00
> and we required three digits of precision so I figured
> that FP was the easiest way to go. I may have been wrong
> there (it being the easiest) but it worked for me.

Obviously I know, I was trying to fish for additional use cases... :-p

How about this: Given a choice between a 24-bit float format 
that's made simple for 8-bit MCUs (e.g. add back the first missing 
bit-1), and 16-bit standards-based half float (though the main 
users are GPUs), which one would you prefer to use?

-- 
Cheers,
Kein-Hong Man (esq.)
Kuala Lumpur, Malaysia


------------------------------------------------------------------------------
Find and fix application performance issues faster with Applications Manager
Applications Manager provides deep performance insights into multiple tiers of
your business applications. It resolves application problems quickly and
reduces your MTTR. Get your free trial!
https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
_______________________________________________
Sdcc-user mailing list
Sdcc-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sdcc-user

Reply via email to