Alan Grimes wrote:

> It would be invisable to the user. It may require some relatively minor 
> changes to the image, I don't know.
> 
> The actual thread implementation will depend on the threading library 
> available on the platform.
> 
> Currently Squeak uses a Many:1 scheduler. It manages "many" processes 
> and schedules them onto one interpreter. I propose to implement a 
> Many:Many scheduler. There would be N interpreters for the number of 
> available computing elements. In addition to these there would be 
> "helper" threads that would be dedicated to IO tasks.

Did you look at a continuation-based model of threads ?


-- 
Serge Stinckwich                                         -< )
Université de Caen>CNRS UMR 6072>GREYC>MAD                /~\
http://purl.org/net/SergeStinckwich                      (/  |
Smalltalkers do: [:it | All with: Class, (And love: it)] _|_/