budd@bu-it.BU.EDU (05/21/89)
Here are the papers I have on the "Arizona" register transfer based
backend technique (used by GNU CC). The first two are well worth
reading. The techniques are so simple, yet so effective I was
flabbergasted when I first read them!
Davidson and Fraser; The Design and Application of a Retargetable
Peephole Optimizer; ToPLaS v2(2) 191-202 (April 1980)
Davidson and Fraser; Register Allocation and Exhaustive Peephole
Optimization; SP&E v14(9) 857-865 (September 1984)
Davidson and Whalley; Quick Compilers Using Peephole Optimization;
SP&E v9(1)? 78-97 (January 1989)
I have heard there is another good paper more recent than
1980/1984. The January 1989 paper refers to;
Davidson and Fraser; Automatic inference and fast interpretation
of peephole optimization rules; SP&E v17 801-812 (1987)
and
Davidson and Fraser; Code selection through object code optimization;
ToPLaS v6 7-32 (1984)
neither of which I have read.
They have a rather minor note (in the Peepholes section) in the
Computing Surveys article (v14n4) by Ganapathi, Fischer and Hennessy,
much as Gries' book contains a small note that Prof. Knuth has been
working on LR(k) grammars, which could turn out to be useful(!)
Also from UofA is David Hanson's work. See; Simple Code Optimizations;
SP&E v13 pp743-763 (1983)
Phil Budne, Boston University
[From budd@bu-it.BU.EDU]
--
Send compilers articles to compilers@ima.isc.com or, perhaps, Levine@YALE.EDU
Plausible paths are { decvax | harvard | yale | bbn}!ima
Please send responses to the originator of the message -- I cannot forward
mail accidentally sent back to compilers. Meta-mail to ima!compilers-request