Related articles |
---|
Basic Blocks bdusan@esu.edu (1996-03-06) |
Re: Basic Blocks ashish@usl.edu (Ashish Ashtekar) (1996-03-08) |
Re: Basic Blocks vinay@uunet.uu.net (1996-03-14) |
From: | vinay@uunet.uu.net (Vinay Kumar) |
Newsgroups: | comp.compilers |
Date: | 14 Mar 1996 17:27:11 -0500 |
Organization: | Compilers Central |
References: | 96-03-046 96-03-071 |
Keywords: | C++, analysis |
>This might sound like a wierd suggestion. The stanford University has
>a compiler group which wrote a compiler sytem called SUIF. One of the
>objective of this system is to provide the user a way to optimize code
>( C or Fortran). The system has a kernel which stores the program's
>intermediate rep. & has C++ interface functions to manipulate the data
>in the kernel.
>
>But you have to look for some way to convert C++ to C, which can be
>passed to the SUIF system.
Hi
I am sorry for jumping in on this so late. Yes, there is a SUIF system
developed by Prof. Monica Lam's group. In fact, we were using it this
quarter to test and implement various optimization passes. The first
thing we had to do was to generate basic blocks and the flowgraph for
the procedure (program).
If I am not mistaken, the code for SUIF is publicly available. As for
converting from C++ to C, someone more knowledge on the net has to
field that one.
In fact, I have the code (written in C++) which identifies the basic
blocks and constructs the flowgraph structure provided the input is in
the SUIF form. The SUIF form itself is a three-address format
imitating many RISC and CISC'y kind of operands.
Send me an e-mail and if you are interested. I will put you in touch
with more knowledgeable people ...
Vinay
vinay@cs.stanford.edu
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.