diff options
author | Sergey Sharybin <sergey.vfx@gmail.com> | 2018-01-08 14:13:56 +0300 |
---|---|---|
committer | Sergey Sharybin <sergey.vfx@gmail.com> | 2018-01-09 18:09:33 +0300 |
commit | 8cffb0a141af6dbfebc6917f2633ab92e26c4d8e (patch) | |
tree | 475a614af44173ccd3914bb43076004b37627a65 /source/blender/blenlib/intern/task.c | |
parent | c4e42d70a4949352f1233574cfc2da30c097439d (diff) |
Task scheduler: Avoid over-allocation of tasks for parallel ranges
This seems to only cause extra rthreading overhead on systems with 10s of
threads, without actually solving anything.
Diffstat (limited to 'source/blender/blenlib/intern/task.c')
-rw-r--r-- | source/blender/blenlib/intern/task.c | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/source/blender/blenlib/intern/task.c b/source/blender/blenlib/intern/task.c index ba600be870b..8c83fcabf79 100644 --- a/source/blender/blenlib/intern/task.c +++ b/source/blender/blenlib/intern/task.c @@ -1104,7 +1104,7 @@ void BLI_task_parallel_range(int start, int stop, * and instead have tasks which are evenly distributed across CPU cores and * pull next iter to be crunched using the queue. */ - num_tasks = num_threads * 2; + num_tasks = num_threads + 2; state.start = start; state.stop = stop; @@ -1257,7 +1257,7 @@ void BLI_task_parallel_listbase( * and instead have tasks which are evenly distributed across CPU cores and * pull next iter to be crunched using the queue. */ - num_tasks = num_threads * 2; + num_tasks = num_threads + 2; state.index = 0; state.link = listbase->first; @@ -1345,7 +1345,7 @@ void BLI_task_parallel_mempool( * and instead have tasks which are evenly distributed across CPU cores and * pull next item to be crunched using the threaded-aware BLI_mempool_iter. */ - num_tasks = num_threads * 2; + num_tasks = num_threads + 2; state.userdata = userdata; state.func = func; |