DSPRelated.com
Forums
Online · Live + on demand
February 9-11, 2027
Price goes up Jan 1: register now to lock in today's rate
Register Now →

ADSP 210XX

Started by Steve Holle ●November 20, 2003
What's special about implementing floating point
calculations in C on a fixed point DSP? Do I have to
use a floating point emulator library or something
like that? Do I have to worry about overflow
truncation etc. all the nasty stuff?

Thanks!

hh_ca

__________________________________


I wasn't even able to import the project.

> -----Ursprungliche Nachricht-----
> oups.yahoo.com]Im Auftrag von Kenneth Porter
> Betreff: Re: AW: [adsp] ADSP 210XX
>
> <> wrote:
>
> > Currently run VDSP 2.0 for our 065 project
> > and VDSP 3.0 for the 161 one. Had no luck
> > porting the 065 project to 3.0 last time. In my
> > early version the handling of makefiles failed.
> > Has anyone had more luck with it?
>
> Our 65L projects work fine with 3.0. We define an environment variable,
> ADI21K, with the value being the date of a compiler release (eg. 20031020
> would be October 20, 2003). We then test this macro in our makefiles to
> perform any release-specific things. It can also be passed as a macro to
> the compiler and linker to work around release bugs and syntax changes.
> (It's particularly useful in LDF files, which change a lot from
> release to
> release.) The important thing is to change the macro's value whenever you
> upgrade the tools, so that the makefiles can make the right decisions.
>