Return to the
comp.compilers page.
Search the
comp.compilers archives again.
| Related articles |
|---|
| Intermediate Representations tony@seagull.rtd.com (Tony Jones) (1994-12-06) |
| Re: Intermediate Representations danhicks@aol.com (1994-12-11) |
| Re: Intermediate Representations hugh@cosc.canterbury.ac.nz (1994-12-14) |
| Newsgroups: | comp.compilers |
| From: | danhicks@aol.com (DanHicks) |
| Keywords: | optimize |
| Organization: | America Online, Inc. (1-800-827-6364) |
| References: | 94-12-056 |
| Date: | Sun, 11 Dec 1994 03:50:08 GMT |
| Received-Date: | Sat, 10 Dec 94 22:50:10 -0500 |
Tony Jones <tony@seagull.rtd.com> writes:
>>>
Can anyone provide references to papers/other sources which compare and
contrast various forms of Intemediate Representations (SSA, PDG, QUADS,
AST's + any others).
I'd prefer real comparisons, rather than references to papers on each IR
(i.e the Ferrante paper on PDG's)
My interest here, is in choosing an flexible but practical IR for a home
project.
<<<
Well, I don't know of any papers, but I can give you some advice: Use a
stack-oriented IR similar to P-code. This is a natural representation to
generate out of a compiler and a natural one to input into a back end.
Most of the books will talk about tuples or elegant things like the Ada IR
(what was it called?) but these are not well suited for **intermediate**
representations. Note that, on the other hand, you will probably want
some sort of tuple representation for an **internal** representation in
the back end.
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.