It was caused by image threading safe commit and it was noticeable
only on really multi-core CPU (like dual-socket Xeon stations), was
not visible on core i7 machine.
The reason of slowdown was spinlock around image buffer referencing,
which lead to lots of cores waiting for single core and using image
buffer after it was referenced was not so much longer than doing
reference itself.
The most clear solution here seemed to be introducing Image Pool
which will contain list of loaded and referenced image buffers, so
all threads could skip lock if the pool is used for reading only.
Lock only needed in cases when buffer for requested image user is
missing in the pool. This lock will happen only once per image so
overall amount of locks is much less that it was before.
To operate with pool:
- BKE_image_pool_new() creates new pool
- BKE_image_pool_free() destroys pool and dereferences all image
buffers which were loaded to it
- BKE_image_pool_acquire_ibuf() returns image buffer for given
image and user. Pool could be NULL and in this case fallback to
BKE_image_acquire_ibuf will happen.
This helps to avoid lots to if(poll) checks in image sampling
code.
- BKE_image_pool_release_ibuf releases image buffer. In fact, it
will only do something if pool is NULL, in all other case it'll
equal to DoNothing operation.
to draw with procedural textures on surface of object. 2D painting will
still paint as if tiled.
When we unify the paint systems, the texture sampling functions will
need to be changed. Sculpt uses a slightly different system that passes
both screen and 3d coordinates to the sampling function. This commit
however is not too disrupting for that however so it can go in now.
After deleting a face, check for unused edges and delete them.
Fixes bug [#33891]
projects.blender.org/tracker/?func=detail&aid=33891&group_id=9&atid=498
Fixes the sculpt object being incorrectly clipped during drawing due
to an out-of-date bounding box making it seem that the object had gone
outside the view.
Added a BKE_pbvh function to get the top-level bounding box. In
sculpt_flush_update(), where the PBVH bounds are updated, the result
is copied to the object's bounding box.
Fixes bug [#33790]
projects.blender.org/tracker/?func=detail&aid=33790&group_id=9&atid=498
Moved the GPU function gpu_bmesh_face_visible() to BKE_paint and
inverted the test to match equivalent tests for other mesh types:
paint_is_bmesh_face_hidden().
Changed BKE_pbvh_bmesh_node_save_orig() to not save hidden faces into
the triangle array.
Modified the non-use-original branch of pbvh_bmesh_node_raycast() to
skip hidden faces.
Fixes bug #33914:
projects.blender.org/tracker/index.php?func=detail&aid=33914&group_id=9&atid=498
Made small tweaks which now limits blur width in a way it can not
be wider than distance between frame boundary and wipe effect center.
It eliminates weirdeness of how blur appears/disappears in cases
effect fac is close to 0/1.
Could be tweaked further but will require effect rewrite it seems.
- BKE_camera_params_from_object now sets the lens to 35 if the object isnt a camera/lamp, without this the view3d lens would be used in camera view.
- running smooth-view multiple times now stores the initial user view rather then over-writing from the current state each time re-smoothview runs.
- redraw panels after smoothview finishes.
- no need for bm_triangle_other_vert_find(), the other vert is known relative to the loop.
- no need for bm_face_edge_backwards(), the verts are in order in the loop.
This adds support of movie textures for Cycles rendering.
Uses the same builtin images routines as packed/generated images,
but with some extra non-rna hookups from blender_session side.
Basically, it's not so clear how to give access to video frames
via C++ RNA -- it'll require exposing ImBuf to API, doing some
threading locks and so. Ended up adding two more functions which
are actually bad level call, but don't consider it's so much bad
-- we have few bad calls already, which are actually related.
Changed a bit how builtin images names are passing to image
manager. Now it's not just an ID datablock name, but also a frame
number concatenated via '@' character, which makes itpossible to
easily know frame number to be used for movie images, without
adding extra descriptors to image manager.
Decoding of builtin name is a bit slower now, but it should be
still nothing in comparison with rendering complexity.
Also exposed image user's frame_current to python API, which
is needed to get absolute frame number of movie from node's
image user.
P.S. Generated/packed images are also using bad level call but
only does it to make things more clear here. Either all images
are using C++ RNA here or no images does. That's the most clear
for now.
Old fixes, since 2009 and before! Related to Curve object editmode:
- normals are not being drawn for hidden curves anymore
- even worse: Blender also didn't hide the extrusions or bevels for hidden curves
- outside edit mode, it shows all (as for all other modes)
- (de)select-all now works when 1st (or any) handle was hidden.
- Previous commit had a line of code of WIP project in it... tss
- Packed Library data was not freed
- Removed recent (post 2.65a) code to exclude packed data from Undo system.
It's internal data thats part of file spec, errors happen for example on
restoring quit.blend or temp saves.
OK - so you have this nice crisp screen, and still you want to add extra
monitors to the laptop! That means Blender should switch back and forth to HiDPI
modes, when you move a window to another monitor.
This code makes the pixelsize scale factor a window property, and handles
an event when a window moves to another monitor. It then changes the
native pixelsize nicely and refreshes entire UI.
You can also have one Blender window on high, and other on low resolution.
Stretching a Blender window from 1 monitor to the other works too, but that
is Apple magic handling it.
Patch makes it possible to compile blender with recent ffmpeg
and libav libraries, mainly by getting rid of deprecated API.
Original patch by Campbell Barton with own modifications to
support compilation with older ffmpeg versions.
This patch could break compatibility of FFV1 videos playing
back in older players, mainly because of alpha support changes.
Preserving compatibility with such players became a headache
and think it's high time to get rid of workarounds here.
The classical SPH solver was not in 2.65, so this change is unlikely to affect many users. But beta users who have been trying it out will need to change the stiffness parameter to sqrt(old value).
Many depsgraph failures are because some data in the graph is being
recalculated too early (or not at all).
Since we better support animators with working renders, here's a hack to
allow manual additional updates on frame changes.
In Property Editor, Object, Panel "Relations Extra" you now have two
buttons:
- Extra Object Update
- Extra Data Update
This will do an extra update of object and/or its data ONLY on frame changes.
Update happens as last.
Tested on files collected in Wiki todo, several cases now work OK, especially
the lags on updates.
