>>>>> "Andreas" == Andreas Doering <[EMAIL PROTECTED]> writes:
Andreas> I guess that an intensive investigation of hardware
Andreas> support for functional programmed systems will show that
Andreas> there is very little that could be added in hardware. And
Andreas> that this little thing does not yield very much.
A couple interesting features that could better support modern
languages were recently discussed on Comp.arch.
One is support for multiple precision arithmetic using larger
digits. Current CPU designs are apparently not as supportive as they
could be because they lack a "carry" flag. The issue I recall is how
to effectively use a "carry" in a superpipeline instruction stream, or
perhaps how to effectively implement the flag without widening
registers, adding more wires, etc.
Another interesting suggestion from Henry Baker was to support
generating all kinds of constants on the CPU in the instruction
stream. This would eliminate the need to go off the CPU to fetch a
constant in RAM (or disk! or in the Age of Java and the Web, a remote
server!)
--
Patrick Logan mailto:[EMAIL PROTECTED]
Voice 503-533-3365 Fax 503-629-8556
Gemstone Systems, Inc http://www.gemstone.com
"We have no intention of shipping another bloated OS and shoving it
down the throats of our users," -Paul Maritz, Microsoft, July 28, 1997
