diff options
| author | Richard Biener <rguenther@suse.de> | 2017-11-17 13:15:34 +0000 |
|---|---|---|
| committer | Richard Biener <rguenth@gcc.gnu.org> | 2017-11-17 13:15:34 +0000 |
| commit | a851ce04f7050dd82aa8344e7b68ee8319fb7b6f (patch) | |
| tree | 4a4b34f1bfe2b59fb26c4e2471c1fff2655af4f4 /gcc/doc/invoke.texi | |
| parent | e716496a557f48cfdf7adb35f5be4e0945bd66e8 (diff) | |
| download | gcc-a851ce04f7050dd82aa8344e7b68ee8319fb7b6f.zip gcc-a851ce04f7050dd82aa8344e7b68ee8319fb7b6f.tar.gz gcc-a851ce04f7050dd82aa8344e7b68ee8319fb7b6f.tar.bz2 | |
re PR fortran/83017 (DO CONCURRENT not parallelizing)
2017-11-17 Richard Biener <rguenther@suse.de>
PR tree-optimization/83017
* tree-parloops.c (MIN_PER_THREAD): Use --param parloops-min-per-thread.
(gen_parallel_loop): Properly count iterations.
(parallelize_loops): Handle loop->can_be_parallel independent
of flag_loop_parallelize_all. Make static profitability test match
the runtime one.
* params.def (PARAM_PARLOOPS_MIN_PER_THREAD): New.
* invoke.texi (parloops-min-per-thread): Document.
* gcc.dg/autopar/pr49960.c: Adjust.
From-SVN: r254867
Diffstat (limited to 'gcc/doc/invoke.texi')
| -rw-r--r-- | gcc/doc/invoke.texi | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi index 1e2b869..e18fa54 100644 --- a/gcc/doc/invoke.texi +++ b/gcc/doc/invoke.texi @@ -10816,6 +10816,12 @@ is 0. Schedule type of omp schedule for loops parallelized by parloops (static, dynamic, guided, auto, runtime). The default is static. +@item parloops-min-per-thread +The minimum number of iterations per thread of an innermost parallelized +loop for which the parallelized variant is prefered over the single threaded +one. The default is 100. Note that for a parallelized loop nest the +minimum number of iterations of the outermost loop per thread is two. + @item max-ssa-name-query-depth Maximum depth of recursion when querying properties of SSA names in things like fold routines. One level of recursion corresponds to following a |