[rescue] Re: sunpci note: TWO cards now available

[Dave McGuire]

On January 30, Big Endian wrote:
> How much faster are the new mentec -11s than the last DEC -11s?  What 
> could they be compared to in terms of more modern vaxen or such?

  I don't have any specs...

> >   Now, the VAX architecture is immensely more complex than the PDP11,
> >but I would think it'd be doable.
> 
> its been done before with charon-vax hasn't it?

  I'm talking about a *hardware* implementation.

> This is sad.  DECpaq now that they've announced the EOL for the Alpha 
> and have discontinued the VAX long before that have dropped *ALL* 
> their nice processor technology and are just Intel whores now.

  Yup.

> How is CISC a limitation?  CISC is merely a design decision (place 
> the burden on the compiler or the processor) isn't it?  the IEEE 

  You're correct of course...I certainly wouldn't be a *technical*
limitation; perhaps I should've been more clear about that.  It'd
likely be a marketing issue.

> floating point should be relatively trivial to add if one designs a 
> new VAX processor.  Some of the higher end vaxen used to have vector 
> processors added to them, these could be integrated ala the G4.  What 
> about scaling to 64 or 128bit?

  I agree.  I haven't looked at how the VAX architecture would scale
width-wise, but I suspect it wouldn't entail too many violations of
the original architecture.  The trick would be maintaining
compatibility.

  But let's face it...32 bits is just fine for most, if not all, of
today's applications.  Granted 64 or 128 bits is MORE, and MORE is
BETTER, but how many times have YOU used integers larger than 2^32 in
your code?  I've done a lot of scientific programming...and *I* sure
haven't.

      -Dave

-- 
Dave McGuire
St. Petersburg, FL         "Less talk.  More synthohol." --Lt. Worf