Hi, just to start with this:
the ZX Spectrum native float format is 5 bytes long.
As is the case for the Amstrad CPC.
For me (interested in Z80 mostly) a 2-byte float makes no sense.
I'd only be interested in a faster float perhaps, then the choice would be 2+1
byte.
but what it'd really appreciate would be the option to use the native float
format of a platform
or a 3rd party library, likely implemented in assembler.
Support for the native format could probably be supplied as a 3rd party library.
This results in 2 requirements:
1. sdcc must be able to adapt it's float moving code to arbitrary sizes
2. sdcc must get a method to encode float literals.
1. would be a little bit nasty to implement probably, as i expect it will break
existing code.
but it's mostly the same work as required for implementing any other float
size.
2. could be done by an included header file which defines float size and bit
assignment.
Then it would be completely up to the programmer to decide, how big her floats
should be.
sdcc could, and should, still provide it's own 32 bit library.
... Kio !
--
http://k1.spdns.de/Vintage/Sinclair/
GPG-ID: 80E6A1DA
signature.asc
Description: Message signed with OpenPGP using GPGMail
------------------------------------------------------------------------------ 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
