Various Topics on Programming

..tri..

> Now, that is not to say that there is no benefit from inline ***embly
> using SIMD or VLIW instructions. Just that these cases are very rare
> and will be uncovered by profiling.
> Since you are not interested in improving the algorithm, you have no
> hope at all for big gains. Any sort of gain you get will be very
> modest (by a small, linear factor) and come at a price of portabilty.

Thanks for your opinions. So then I wonder if anybody could suggest
me a way figure out how to best figure out if SIMD or VLIW is a good
approach
for a specific application. Is there any helpful
tool out there? My other approach was doing it in someway as "manual
work".
Using a cycle accurate simulator and see what instructions are
executed
and then analysing what could be done in parallel. Is there a better
way to
do that?

Cheers,
Patrick

..tri..

me again . Was thinking of some kind of dataflow programm that
automatically generates the
dataflow from a C application. Is something like this around?

CHeers