"Andreas Raab" <andreas.raab at gmx.de> wrote:
[snip]
> 
> Let's play this fair, shall we? If we do optimizations that are specific for
> a single platform, let's move it where it belongs - into a macro living
> inside the platform specific code with the default being something that we
> can expect to work reasonably across all platforms, not just our
> pet-platform.
Dang! I was about to write almost exactly what you said, apart from the
details of Mac compilers (which I suspected but didn't know).

It's exactly why  I recently factored out the direct-jump primitive
calling to a macro. I think I succesfully made it so one could still go
via a switch if truly required.

Getting too close to a particular cpu (version)/ architecture is just as
bad as early binding in the higher language. A new version of the
nominally same cpu can be effectively a new port - forget the gulf
between architectures. Where possible we need to produce good 'normal'
C that a decent compiler will do well with. In places we may want to
provide for a manual override (I have had cases where a trivial asm
routine can replace complicated C for big improvements) for some
machines.

Summary: <aol>Me too</aol>

tim
--
Tim Rowledge, tim at sumeru.stanford.edu, http://sumeru.stanford.edu/tim
Strange OpCodes: IKI: Ignore Keyboard Input