Re: fast table generators

goer@midway.uchicago.edu (Richard L. Goerwitz) writes:
>[I've written LR(1)-family table generators, they always run slow.
> Any practical implementation books?]


If you're looking for big-O speedeups, I don't know, but if you're looking
for clever implementations, consider merged coded and data in the tables.
(The first paper is relevant because the code generator is a parser).


3 Citations Follow:
%A Chris W. Fraser
%A Robert R. Henry
%T Hard-Coding Bottom-Up Code Generation Tables to Save Time and Space
%J Software \- Practice and Experience
%V 21
%N 1
%D January 1991
%P 1-12


%A Frank G. Pagan
%T Comparative Efficiency of General and Residual Parsers
%J SIGPLAN Notices
%V 25
%N 4
%D April 1990
%P 59-68
%X * Focus on partial evaluation of parsers generating pure code.
  * Hand traslation of Pascal and C.
  * Time (2-10X faster) and size (0.5-1.25 larger).
  * View as explicit partial evaluation of parser wrt to grammar
(doesn't state that it is also partial evaluation of interpreter wrt a
graph).


%A Thomas J. Pennello
%T Very Fast LR Parsing
%J Proceedings of the SIGPLAN 1986 Symposium on Compiler Construction;
SIGPLAN Notices
%V 21
%N 7
%D July 1986
%P 145-151
%X * Partial evaluation of the table interpreter with resepct to each
element of the table (not presented as such).
  * Speedup: on a VAX-like machine, 40,000 to 500,000 lines per minute.
On an 80286, 37,000 to 240,000 lines per minute.
  * FSM converted to assembly language.
  * 2-4X increase in table size.


Hope these help.


        ;-D oN ( "You got code in my data." "You got data in my code!" ) Pardo
--

Return to the comp.compilers page.
Search the comp.compilers archives again.