diff options
| author | Diego Novillo <dnovillo@gcc.gnu.org> | 2006-01-23 15:03:24 -0500 |
|---|---|---|
| committer | Diego Novillo <dnovillo@gcc.gnu.org> | 2006-01-23 15:03:24 -0500 |
| commit | 7ebaeab50e26463a6fbe6fe23cffdf69a84bce30 (patch) | |
| tree | 62676ee57198338096184e56a91032bf1d280983 | |
| parent | 8a0092c83488d75b6733d741804b5d961d5f9289 (diff) | |
| download | gcc-7ebaeab50e26463a6fbe6fe23cffdf69a84bce30.zip gcc-7ebaeab50e26463a6fbe6fe23cffdf69a84bce30.tar.gz gcc-7ebaeab50e26463a6fbe6fe23cffdf69a84bce30.tar.bz2 | |
omp-low.c: Update documentation.
* omp-low.c: Update documentation.
From-SVN: r110143
| -rw-r--r-- | gcc/ChangeLog | 6 | ||||
| -rw-r--r-- | gcc/omp-low.c | 11 |
2 files changed, 8 insertions, 9 deletions
diff --git a/gcc/ChangeLog b/gcc/ChangeLog index 2a85326..0c2c7ba 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,5 +1,9 @@ +2006-01-23 Diego Novillo <dnovillo@redhat.com> + + * omp-low.c: Update documentation. + 2006-01-23 Andrew Pinski <pinskia@physics.uc.edu> - Daniel Berlin <dberlin@dberlin.org> + Daniel Berlin <dberlin@dberlin.org> PR tree-opt/25315 PR tree-opt/25857 diff --git a/gcc/omp-low.c b/gcc/omp-low.c index f5bdcb9..44f1b7e 100644 --- a/gcc/omp-low.c +++ b/gcc/omp-low.c @@ -50,14 +50,9 @@ Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA re-gimplifing things when variables have been replaced with complex expressions. - Lowering of a parallel statement results in the contents of the - parallel being moved to a new function, to be invoked by the thread - library. The variable remapping process is complex enough that only - one level of parallel statement is handled at one time. If there are - nested parallel statements, those nested statements are handled when - the new function is lowered and optimized. The result is not 100% - optimal, but lexically nested parallels effectively only happens in - test suites. */ + Final code generation is done by pass_expand_omp. The flowgraph is + scanned for parallel regions which are then moved to a new + function, to be invoked by the thread library. */ /* Parallel region information. Every parallel and workshare directive is enclosed between two markers, the OMP_* directive |