Comparing Control Flow Graphs

Is there any formal notion that captures the "precision" of a control flow
graph. For example, assume we have a source program M and two intermediate
representations M_a and M_b. The control flow graphs for M_a is C_a and
the control flow graph for M_b is C_b. Do people have a formal way of
saying that C_a is "better" than C_b or vice-versa. The notion of "better"
should imply better solutions to some class of dataflow problems. Here is
a contrived example.


Main program:


procedure Z
if (x = y) then A else B;
C;
end




Control Flow Graphs:


x = y x = y
/\ /\
/ \ / \
A B A;C B;C
\ / \ /
\/ \/
| |
C |
| |
end end


--
Amr Sabry sabry@cs.rice.edu
--

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