Call for 4.3 project reviewers

[Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>]

Hello,

in accordance with http://gcc.gnu.org/ml/gcc/2006-09/msg00454.html, I am
looking for reviewers for my 4.3 projects.  I would like to get someone
"responsible" for reviewing each project soon, so that I may discuss the
changes he considers necessary with him.  For each project, the list of
touched areas of compiler together with estimates of the extent of the
changes is provided (not all the changes necessarily have to be merged
as part of the project, some of them qualify as bugfixes or independent
improvements).

Code for predictive commoning is available on predcom branch.  Regarding the
remaining three projects, at the moment I have them only available locally;
I plan to put them on branch soon (I must first decide whether I want
to create a separate branch for each of them, or just one common):

stage 1:

predictive commoning (http://gcc.gnu.org/wiki/PredictiveCommoning):
  The predictive commoning pass itself (2.5kloc).  Significant changes
  to loop manipulation (unrolling, updating of frequencies).  Uses
  affine expressions framework
  (http://gcc.gnu.org/ml/gcc-patches/2006-03/msg00345.html).
  Changes to dependence analysis (more correct handling of aliasing).
  6.5kloc diff in total.

stage 2:

autoparallelization (http://gcc.gnu.org/wiki/AutomaticParallelization):
  In addition the the autoparallelization pass (1.5kloc), minor changes
  to gomp framework.  2kloc diff in total.

prefetching improvements (http://gcc.gnu.org/wiki/PrefetchingImprovements):
  In addition to the changes to prefetching (tree-ssa-loop-prefetch.c):
  New cross-loop reuse analysis pass (2.5kloc).  Significant changes to
  # of iterations estimation.  Minor changes to rtl expansion (expr.c,
  optabs, ...) to enable generation of non-temporal store instructions.
  Minor changes to loop manipulation functions.  Some cleanups to
  tree-data-ref.c (to enable sharing of code with reuse analysis).
  6.5kloc diff in total.

preserving loop structures during optimizations (http://gcc.gnu.org/wiki/PreservingLoops):
  Major changes to jump threading infrastructure
  (tree-ssa-threadupdate.c) and optimizers using it (dom, vrp),  and
  loop structure updating functions (cfgloop*.c).  Minor changes to many
  tree-ssa optimizers, and pass manager.

Volunteers, please?

Zdenek