[CCC DEV] Floating point on Arduino...

Matt Jadud matt at jadud.com
Wed Jun 9 19:40:21 BST 2010


On Wed, Jun 9, 2010 at 11:55 AM, Adam Sampson <ats at offog.org> wrote:
> If you mean adding the floating-point instructions, it'd mean making it
> emulate a kind of Transputer that never existed, which I would imagine
> would have interesting consequences elsewhere in the toolchain. ;-) I

Aye. I spoke imprecisely. I see your point, though. I forgot we were
running T2 and not T4 on the AVR.

> program from working is that every REAL32 constant in the program gets
> emitted as 0x00000000 by the compiler! I have a vague recollection of
> having seen this problem before -- does it ring bells with any of you
> guys?

No. I can poke around, though.

The target for this work is an Atmega1280, so we'll have a fair bit
more flash to work with (virtual memory implementation, of course,
being a stepping-stone to that extra flash). The UAV project is nice
because it has some real mathematical motivators built into it, so
either getting FP emulation working (if not in the category of the
impossible) or binding to <math.h> in avr-libc are nice additions (in
general) and very necessary in the short-run.

Or, we can convert all of our mathematical operations to integer
operations... which might be tricky.

Thanks for the pointers. I'll let you know what my explorations turn up.

Cheers,
M




More information about the developers mailing list