thinking about virtual machines

Michael Korns mkorns@ix.netcom.com
Mon, 28 Apr 1997 10:04:40 -0700


Fare,

> > We need a more machine centric VM which is easy to translate into
native
> > binary on disparate machines.
> >
> That I'm not convinced of at all.
> Centric to what machine?
> For instance what about clusters of MISC computers?
> They'd overpower any stupid CISC/RISC bloatcessor for epsilon of the
price.
> but if they have to emulate a CISC/RISC bloatcessor, they will suck.
> Have you really studied the cost of the various compiler passes,
> and wouldn't keeping things high-level (including the generic
> high-level code analyses) keep things much more portable and efficient?
> [Again, using FAT precompiled code for the 2% critical routines].
> What about restrict portability to stubborn von Neuman style machines?
> The von Neuman model is clearly showing its limits performance/cost-wise!

I am in agreement that a High Level VM (iff it could be made efficient in
portable native code generation) could be better than a 32 bit "quad" VM.
However, we have working examples (in AgentBase for instance) of 32 bit
VM's being fast and efficient. We would need practical research and
examples of efficient High Level VM's. For instance, I'm looking into the
Juice literature.