On Thu, Mar 7, 2013 at 5:39 PM, David Rowe <da...@rowetel.com> wrote:

> On a lighter note I once wrote a "fixed point fairy tale" to vent my
> frustration at this same problem in a corporate situation, where I've
> seen it derail companies, cost jobs and millions of $:
>
>         http://www.rowetel.com/blog/?p=10

After building the fixed-point C version, you may be able
to try it out on the DSP. It may be good enough. If not,
you have compiler-generated assembly to start from; you
will not need to worry about doing worse than a compiler.

An alternative approach, especially if you are only targeting
one assembly language, is to process that assembly back
into C for testing. It's often roughly a matter of a sed script.
Registers turn into local variables, jumps turn into goto, etc.
Transformations of the opcodes can even add profiling code,
counting in a way that matches instruction cost on the target.

------------------------------------------------------------------------------
Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester  
Wave(TM): Endpoint Security, Q1 2013 and "remains a good choice" in the  
endpoint security space. For insight on selecting the right partner to 
tackle endpoint security challenges, access the full report. 
http://p.sf.net/sfu/symantec-dev2dev
_______________________________________________
Freetel-codec2 mailing list
Freetel-codec2@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freetel-codec2

Reply via email to