Multy-core CPUs

Igor Stasenko siguctua at gmail.com
Thu Oct 25 21:08:51 UTC 2007


Too much to read and comment i wish to have time for all of this..

If we have different models for solving different concurrency problems
- then its obviously must be at language side, not VM side. Enforcing
a particular solution for VM will render other solutions hard or even
impossible to implement.
And i agree, while Erlang provides a solution for Erlang, its clearly
not the silver bullet for Smalltalk.

As for VM: i think we have to deal with parallelism and concurrency
problems of VM alone (change memory management and interpreter to
support a number of native threads running in parallel). Then rest
should be considered at language side.

Then we will stop arguing and implement a libraries each proposing own
good/bad solutions to concurrency and be happy :)


-- 
Best regards,
Igor Stasenko AKA sig.



More information about the Squeak-dev mailing list