RE: Strongtalk and Ruby

> -----Original Message-----
> From:  Richard
>
> Does anyone know how suited the ruby language is to run on the
> strongtalk VM?

Search the list for past messages about Ruby.  This has been discussed
before.  Ruby should be an excellent target for the Strongtalk virtual
machine.  However, Ruby has a very messed-up grammar, which makes it hard to
write a new parser for it.

> I see that the ruby community has a couple of efforts
> to develop virtual machines (jruby and yarv), but I guess so far these
> have not lead to much of a speed increase, and it seems that they
> don't have the techniques employed in the strong talk VM. Being a
> rubyist myself, I often miss being able to specify types and have the
> feeling that ruby would benefit significantly having optional static
> typing. Hence a second argument for wanting to target the stongtalk
> VM.

Strongtalk's type system has nothing to do with the virtual machine.
Porting Ruby to the Strongtalk VM would not give Ruby a type system.

> [...]
>
> What is the situation with threads and strongtalk? Is there any chance
> that strongtalk will have any more than green threads? Does it have
> green threads (os runs one thread, vm jumps between green threads,
> uses non-blocking io)?

Strongtalk does not use green threads.  It uses native threads.  However
only one thread is allowed to run at a time, since the VM is not
multithreaded.  However, use of native threads means that Strongtalk can
easily do blocking I/O.  Callouts can be designated as asynchronous, meaning
that other Smalltalk threads can be allowed to proceed while that callout
executes or blocks.

> I hope you'll excuse that little example. As maybe you can tell I much
> prefer that ruby syntax the smalltalk syntax, but what I'm asking is:
> under the covers at the VM level is pretty much all the same? Will
> inlining remove all or most of the block structures in my example?
> Would my ruby code run 10 times faster on the strongtalk VM (there's a
> network call in the middle of the example, so I guess I mean, use 10%
> of the cpu resources, not run 10 times faster. As far as I know the
> ruby interpreter doesn't do any inlining)?

Current Ruby VMs are very, very slow compared to Strongtalk, so the speedup
would probably be a lot more than a factor of 10.  Strongtalk is more than
10 times as fast as Squeak, which is itself faster than Ruby, I believe.
Yes, inlining should probably remove all the blocks in your example;
generally, any block that is only passed downwards as a literal argument to
sends that are themselves inlined, will be inlined.

-Dave


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Strongtalk-general" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [email protected]
For more options, visit this group at http://groups.google.com/group/strongtalk-general?hl=en
-~----------~----~----~----~------~----~------~--~---