[DGD] Current Status of VM efforts

[Carter Cheng]

Thanks Felix,

I think I might make an attempt as a learners effort to understand the
LLVM system to construct some sort of translation layer for code in
the vm system. I do have a quick question though- how does the VM
invalidate programs when they are recompiled and destroyed?

Regards,

Carter.

On Sun, Aug 7, 2011 at 7:11 PM, Felix A. Croes <felix at dworkin.nl> wrote:
> Carter Cheng <cartercheng at gmail.com> wrote:
>
>> I am somewhat curious about the current status of the VM efforts wrt
>> to the Hydra and dgd-osr projects. Is there currently a JIT planned
>> for this?
>
> I just merged my newvm branch, which updates DGD's virtual machine,
> into the main repository.  The old VM was almost 20 years old, and
> getting a bit long in the tooth.  Old code will still work, but newly
> compiled LPC code will run on the new virtual machine.
>
> The old virtual machine has a peculiarity that it shares with the
> other LPmud servers: it has lvalues (left-hand values in an assignment)
> as a runtime type.  The new VM does away with these, bringing it more
> into line with present-day VM designs.  This should also simplify
> compiling code from DGD's VM to a different architecture.
>
> As to a JIT compiler for DGD, all I have is the notion that it should
> compile from the internal VM code to a mainstream virtual machine
> (Java, Mono, LLVM?) and possibly to C, so that it could be architecture-
> independent.  The drawback would be multiple translation stages
> (DGD VM code -> LLVM code -> host machine code).  I have no settled
> plans for starting work on this, so you're certainly welcome to
> give it a try, yourself.
>
> Regards,
> Felix Croes
> ___________________________________________
> https://mail.dworkin.nl/mailman/listinfo/dgd
>