Return to the
comp.compilers page.
Search the
comp.compilers archives again.
| Related articles |
|---|
| re: compiling techniques compilers@ima.UUCP (1986-01-08) |
| Relay-Version: | version B 2.10.2 9/12/84; site mit-hermes.ARPA |
| Posting-Version: | Notesfiles $Revision: 1.6.2.16 $; site ima.UUCP |
| From: | compilers@ima.UUCP |
| Newsgroups: | mod.compilers |
| Date: | 8 Jan 86 16:35:00 GMT |
| Article-I.D.: | ima.136300035 |
| Posted: | Wed Jan 8 11:35:00 1986 |
| Date-Received: | 10 Jan 86 11:22:25 GMT |
| Nf-ID: | #N:ima:136300035:000:856 |
| Nf-From: | ima!compilers Jan 8 11:35:00 1986 |
[from ]
Has anybody seen the paper on "data descriptors" by Holt and Cordy?
It comes with the distribution tape for Concurrent Euclid (from the
University of Toronto). I don't know offhand if it's been published
in any other form.
It describes a uniform internal representation for data objects
(eg. constants, variables, records, addresses, etc). Since the representation
is uniform, it makes life a lot simpler and makes all sorts of optimizations
managable. For example, the Concurrent Euclid compiler (which uses
data descriptors) can easily detect when the target of an assignment
appears in the right hand side and emit better code in such cases.
>From what little I've read about Graham-Glanville code generation,
I'd swear that it would benefit by incorporating data descriptors into
it's grammars.
Paul Tarvydas
{ihnp4,etc.}!utcsri!tarvydas
Return to the
comp.compilers page.
Search the
comp.compilers archives again.