Age | Commit message (Collapse) | Author |
|
|
|
The idea is to accumulate all new tasks in a thread local queue
first without doing any thread synchronization (aka, locks and
conditional variables) and move those tasks to a scheduler queue
once they are all ready. This way we avoid per-task-pool lock
and only have one lock per bunch of tasks.
This is particularly handy when scheduling new dependency graph
node children. Brings FPS of cached simulation from the linked
below file from ~30 to ~50.
See documentation for BLI_task_pool_delayed_push_{begin, end}
and for TaskThreadLocalStorage::do_delayed_push.
Fixes T50027: Rigidbody playback and simulation performance regression with new depsgraph
Thanks Bastien for the review!
|
|
|
|
These bits became obsolete with the new layer system, so we can
simplify some code around them or avoid existing workarounds which
were trying to keep things working for them.
There are still work needed to be done for on_visible_change to
avoid unnecessary updates, but that can also happen later.
|
|
Suspended pools allows to push huge amount of initial tasks
without any threading synchronization and hence overhead.
This gives ~50% speedup of cached rigid body with file from
T50027 and seems to have no negative affect in other scenes
here.
|
|
This is something what should be done in the task scheduler instead
with local thread queues so we handle this in a single place.
|
|
This feature was adding extra complexity to task scheduling
which required yet extra variables to be worried about to be
modified in atomic manner, which resulted in following issues:
- More complex code to maintain, which increases risks of
something going wrong when we modify the code.
- Extra barriers and/or locks during task scheduling, which
causes extra threading overhead.
- Unable to use some other implementation (such as TBB) even for
the comparison tests.
Notes about other changes.
There are two places where we really had to use that limit.
One of them is the single threaded dependency graph. This will
now construct a single-threaded scheduler at evaluation time.
This shouldn't be a problem because it only happens when using
debugging command line arguments and the code simply don't
run in regular Blender operation.
The code seems a bit duplicated here across old and new
depsgraph, but think it's OK since the old depsgraph is already
gone in 2.8 branch and i don't see where else we might want
to use such a single-threaded scheduler.
When/if we'll want to do so, we can move it to a centralized
single-threaded scheduler in threads.c.
OpenGL render was a bit more tricky to port, but basically we
are using conditional variables to wait background thread to
do all the job.
|
|
|
|
This kind of keeps threads "warmer" and should in theory give better
cache coherency bringing some %% of speedup. It was already tested
few months ago and it gave few % speedup in barber shop, but was
reverted due to some bone popping. The popping is now fixed so it
should be fine to use new scheduling policy.
|
|
|
|
This reverts commit 9444cd56db1a4e43d03fa8c735cd893b2e74b913.
This commit caused some flickering in the bones when swapping IK to Fk.
While it's unclear why such change caused any regressions, let's revert
it to unlock the studio.
|
|
The idea here is again to finish objects evaluation as soon as possible.
Seems to be giving another 3% speedup in the barber scenes.
|
|
|
|
|
|
This is mainly a maintenance commit which was aimed to make work with
this module more pleasant and solve such issues as:
- Annoyance with looong files, which had craftload in them
- Usage of STL for the data structures we've got in BLI
- Possible symbol conflicts
- Not real clear layout of what is located where
So in this commit the following changes are done:
- STL is prohibited, it's not really predictable on various compilers,
with our BLI algorithms we can predict things much better.
There are still few usages of std::vector, but that we'll be
solving later once we've got similar thing in BLI.
- Simplify foreach loops, avoid using const_iterator all over the place.
- New directory layout, which is hopefully easier to follow.
- Some files were split, some of them will be split soon.
The idea of this is to split huge functions into own files with
good documentation and everything.
- Removed stuff which was planned for use in the future but was never
finished, tested or anything.
Let's wipe it out for now, and bring back once we really start using
it, so it'll be more clear if it solves our needs.
- All the internal routines were moved to DEG namespace to separate
them better from rest of blender.
Some places now annoyingly using DEG::foo, but that we can olve by
moving some utility functions inside of the namespace.
While working on this we've found some hotspot in updates flush, so
now playback of blenrig is few percent faster (something like 96fps
with previous master and around 99-100fps after this change).
Not saying it's something final, there is still room for cleanup and
API simplification, but those might happen as a regular development
now without doing any global changes.
|