[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [avr-gcc-list] Double desirability
From: |
Joerg Wunsch |
Subject: |
Re: [avr-gcc-list] Double desirability |
Date: |
Thu, 7 Apr 2005 11:39:38 +0200 (MET DST) |
Parthasaradhi Nayani <address@hidden> wrote:
> I am wiilling to develop the code, but I am not sure about the time
> frame as of now. If the current code is in assembly, then I would
> prefer to continue the same in assembly and also try and plug the
> holes.
It's not strictly the requirement I would impose though. The existing
code has proven to be very hard to maintain (and the existing NaN/Inf/
denormal handling bugs are unlikely to be fixable without rewriting a
lot of it), so if a non-assembler implementation would get us to say
110...120 % of the code size of the assembler hacks we've got now, but
with a 500 % maintainability improvement, I wouldn't mourn. Floating
point is certainly not really suitable for an ATtiny2313 anyway.
--
J"org Wunsch Unix support engineer
address@hidden http://www.interface-systems.de/~j/