several messages

Jeffrey A Law law@cygnus.com
Tue Aug 29 18:27:00 GMT 2000


  In message < 10008300122.AA16995@vlsi1.ultra.nyu.edu >you write:
  >     One of the significant problems with GCC is that it (in many places) do
  > es
  >     not use well-understood algorithms, which makes it difficult for people
  >     who do compilers for a living to understand (and ultimately contribute)
  >  to
  >     GCC.
  > 
  > And then you've got people like me who understand the old GCC algorithms
  > completely, but never did know much of anything about the "well-understood"
  > conventional algorithms.  ;-)
:-)  Part of the problem is the "well understood" conventional algorithms
weren't that widely publicized until recently.  One had to pour over
conference proceedings to know what was going on (particularly in the world
of optimizations).

In the last few years there's been 2 good and 1 OK textbook published
(and a 4th on the way which should be good) which complement the Dragon
book well.

The explosion of the web has also increased our ability to retrieve
information that normally was only available to conference attendees
(or those that pay for ACM publications).

Add to those changes in the availability of information our changes to
provide sensible building blocks for modern optimizations and we've got
the opportunity to actually *use* the optimizations we read about.

jeff



More information about the Gcc-patches mailing list