Path: | doc.ic.ac.uk!dds |
From: | dds@doc.ic.ac.uk (Diomidis Spinellis) |
Newsgroups: | comp.lang.functional |
Subject: | Re: Why Abstract Machines ? |
Date: | 26 Oct 1993 13:25:02 GMT |
Organization: | Dept of Computing, Imperial College, England |
Lines: | 22 |
Distribution: | world |
Message-ID: | <2aj8fe$dve@frigate.doc.ic.ac.uk> |
References: | <1993Oct26.092132.4242@cse.iitb.ernet.in> |
NNTP-Posting-Host: | scorch.doc.ic.ac.uk |
Content-Length: | 1084 |
>From article <1993Oct26.092132.4242@cse.iitb.ernet.in>, by oberoi@cse.iitb.ernet.in (Sundeep Oberoi):
[...]
> The process of abstracting away from
> physical machine may take away opportunities to utilize particular
> hardware features. The problem of abstracting away from the language may
> also rob one of some optimization opportunities as the abstraction process
> tend to cause some loss of information.
Certain abstract machine implementations, on certain architectures can
actually be faster than native code compilation schemes. This can be
the case when the abstract machine interpreter and parts of the abstract
machine instructions fit in instruction and data caches that would
otherwise be thrashed by lengthy native code sequences.
In older microcoded CISC architectures abstract machines could be
implemented by a modified microcode intepreted (WAM on a VAX comes
to mind.)
Diomidis
--
Diomidis Spinellis Internet: <dds@doc.ic.ac.uk> UUCP: ...!uknet!icdoc!dds
Department of Computing, Imperial College, London SW7 #include "/dev/tty"
Newsgroup comp.lang.functional contents
Newsgroup list
Diomidis Spinellis home page
Unless otherwise expressly stated, all original material on this page created by Diomidis Spinellis is licensed under a Creative Commons Attribution-Share Alike 3.0 Greece License.