Age | Commit message (Collapse) | Author |
|
|
|
|
|
Was a mistake in one of the previous TLS commits.
See comment in the pool_create to see some details why it was crashing.
|
|
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.
|
|
The idea is to allow some amount of tasks to be pushed from working
thread to it's local queue, so we can acquire some work without doing
whole mutex lock.
This should allow us to remove some hacks from depsgraph which was
added there to keep threads alive.
|
|
This allows us to avoid TLS stored in pool which gives us advantage of
using pre-allocated tasks pool for the pools created from non-main thread.
Even on systems with slow pthread TLS it should not be a problem because
we access it once at a pool construction time. If we want to use this more
often (for example, to get rid of push_from_thread) we'll have to do much
more accurate benchmark.
|
|
Basically move all thread-specific data (currently it's only task
memory pool) from a dedicated array of taskScheduler to TaskThread.
This way we can add more thread-specific data in the future with
less of a hassle.
|
|
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.
|
|
|
|
Initialize a rectangle from point+size.
|
|
Comments said that function was supposed to 'stop worker threads', but
it absolutely did not do anything like that, was merely wiping out TODO
queue of tasks from given pool (kind of subset of what
`BLI_task_pool_cancel()` does).
Misleading, and currently useless, we can always add it back if we need
it some day, but for now we try to simplify that area.
|
|
processed.
Freeing pool was calling `BLI_task_pool_stop()`, which only clears
pool's tasks that are in TODO queue, whithout ensuring no more tasks
from that pool are being processed in worker threads.
This could lead to use-after-free random (and seldom) crashes.
Now use instead `BLI_task_pool_cancel()`, which does waits for all tasks
being processed to finish, before returning.
|
|
|
|
|
|
Not really happy of per-pool threads limit, need to find better
approach to that. But at least it's possible to get rid of half
of the nastyness here by removing getter which was only used in
an assert statement.
That piece of code was already well-tested and this code becomes
obsolete in the new depsgraph and does no longer exists in blender
2.8 branch.
|
|
This was only used for progress report, and it's wrong because:
- Pool might in theory be re-used by different tasks
- We should not make any decision based on scheduling stats
Proper way is to take care of progress by the task itself.
|
|
- remove unused struct member.
- misleading variable name.
|
|
|
|
|
|
|
|
We (the Microsoft C++ team) use the Blender project as part of our "Real world code" tests.
I noticed a place in WIN32 specific code (dvpapi.cpp:85) where a string literal is losing
its const-ness when being passed to BLI_dynlib_open(). This is not permitted when using the
/permissive- conformance compiler switch (see our blog
https://blogs.msdn.microsoft.com/vcblog/2016/11/16/permissive-switch/)
My suggested fix is to add const and propagate it where needed. Another possible fix would be
to explicitly cast away the const.
Reviewers: mont29, sergey, LazyDodo
Subscribers: Blendify, sergey, mont29, LazyDodo
Tags: #platform:_windows
Differential Revision: https://developer.blender.org/D2495
|
|
The new `isect_ray_aabb_v3_simple` function replaces the `BKE_boundbox_ray_hit_check` and can be used in BVHTree Root (first AABB). So it is much more efficient.
|
|
|
|
|
|
in BLI_string_utils.
|
|
Functions like that do not have anything to do in BKE really, even less
when actually more used for bones than vgroups!
|
|
Things like `BLI_uniquename` had nothing, but really nothing to do in
BLI_path_util files!
Also, got rid of length limitation in `BLI_uniquename_cb`, we can use
alloca here to avoid overhead of malloc while keeping free size (within
reasonable limits of course).
|
|
This adds two functions to project 3d coordinates onto a 3d plane,
to get 2d coordinates, essentially eliminating the plane's normal axis
from the coordinates.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D2460
|
|
This splits `interp_weights_face_v3` into `interp_weights_tri_v3` and
`interp_weights_quad_v3`, in order to properly handle three sided polygons
without needing a useless extra index in your weight array. This also
improves clarity and consistency with other math_geom functions, thus
reducing potential future errors.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D2461
|
|
Other than implementing a `mid_v3_v3_array` function, this removes
`cent_tri_v3` and `cent_quad_v3` in favor of `mid_v3_v3v3v3` and
`mid_v3_v3v3v3v3` respectively.
Reviewed By: mont29
Differential Revision: https://developer.blender.org/D2459
|
|
Seems like it was erroring on some buildbots...
|
|
|
|
That code was a joke, letting some invalid utf8 bytes pass, returning
wrong offset for some invalid sequences, not to mention length and
pointer easily going out of sync, NULL final byte being 'forgotten' by
memcpy, etc. etc.
The miracle here is that we could survive using this for so long!
Probably because we do not use utf-8 sanitizing enough in Blender,
actually... :/
|
|
|
|
Note which GL functions these are equivalent to.
|
|
- Expand overly dense & confusing delta assignments.
- Replace bit shift with multiply.
Also link to 'clipped' version of this function
which may be useful to add later.
|
|
Seems to be a typo in recent commit e1e49fd.
|
|
No need to have temp array storage, avoid 2x loops.
|
|
- Remove 'rotate_m2', unlike 'rotate_m4' it created a new matrix
duplicating 'angle_to_mat2' - now used instead.
(better avoid matching functions having different behavior).
- Add 'axis_angle_to_mat4_single',
convenience wrapper for 'axis_angle_to_mat3_single'.
- Replace 'unit_m4(), rotate_m4()' with a single call to 'axis_angle_to_mat4_single'.
|
|
|
|
Existing method was fine for basic polygons but didn't scale well
because its was checking all coordinates for every y-pixel.
Heres an optimized version.
Basic logic remains the same this just maintains an ordered list of intersections,
tracking in-out points, to avoid re-computing every row,
this means sorting is only done once when out of order segments are found,
the segments only need to be re-ordered if they cross each other.
Speedup isn't linear, test with full-screen complex lasso gave 11x speedup.
|
|
|
|
Bitmap drawing is out-of-scope for a general math API,
move to BLI_bitmap_draw_2d.
|
|
Self-explanatory. to find broken links run `sphinx-build -b linkcheck sphinx-in sphinx-out`
Reviewers: mont29
Tags: #bf_blender, #python, #infrastructure:_websites
Differential Revision: https://developer.blender.org/D2297
|
|
We were checking for number of tasks from given pool already active, and
then atomically increasing it if allowed - this is not correct, number
could be increased by another thread between check and atomic op!
Atomic primitives are nice, but you must be very careful with *how* you
use them... Now we atomically increase counter, check result, and if we
end up over max value, abort and decrease counter again.
Spotted by Sergey, thanks!
|
|
Another case of float imprecision leading to endless loop. INcreasing a bit 'noise threashold' seems to work OK.
Not a regression, but might be nice to have in 2.78a.
|
|
This makes cross compilation a little less painful
|
|
This was logically incorrect and was causing warning and compilation
errors with strict compiler flags.
|
|
Was using wrong argument name in doxygen comment. Also reduced scope of vars.
Sorry for the noise :/
|
|
We were calling BLI_remlink and then BLI_insertlinkbefore/after quite often. BLI_listbase_link_move simplifies code a bit and makes it easier to follow. It also returns if link position has changed which can be used to avoid unnecessary updates.
Added it to a number of list reorder operators for now and made use of return value. Behavior shouldn't be changed.
Also some minor cleanup.
|