Creates a projection matrix for a small region of the viewport.
Reviewers: campbellbarton, brecht
Differential Revision: https://developer.blender.org/D5412
Previously we were setting it to 1 (aka no 'chunking'), to follow
previous behavior. However, this is far from optimal, especially with
CPUs that can have tens of threads nowadays.
Now taking an heuristic approach (inspired from the one already existing
for `BLI_task_parallel_listbase()`, which tries to guesstimate best
chunk sizes based on several factors (amount of threads/parallel tasks,
total number of items, ...).
Think this is a reasonable base ground, more optimization here would of
course be possible.
Note that code that was already explicitely settings some value here
won't be affected at all by that change.
Function would not clear the static scheduler pointer, which lead to
crash (mem use after free) when trying to re-init and use the task API
again. Should not happen in Blender itself, but could in other cases
(like some future gtests ;) ).
Some versions of the Windows SDK headers have a dbghelp.h that will emit
C4091 warnings, repress them just this once, since the warn can be helpful
in other places.
This patch makes BLI_task_scheduler_create wait for all worker threads to have started before
returning to caller. For very short workloads (BLI_taks_test) there is the chance that the
worker threads have not fully started yet, and the main thread is calling pthread_join at
the same time as pthread_setspecific is being called on the worker threads which causes a
deadlock on pthreads4w.
Differential Revision: https://developer.blender.org/D4936
Reviewed By: mont29, sergey, brecht
The declaration and implementation of BLI_path_name_at_index were
out of sync leading to build warning
C4028: formal parameter 1/3/4 different from declaration
Remove the clear allocation flag as it has little impact since there should
be very few allocation per redraw.
Make BLI_memblock_alloc and BLI_memblock_iterstep much more cache efficient
removing them almost entirely from performance profiles.
very straightforward: initialize default to the same hard limit as the
RNA properties.
Annoying part is that it's not trivial to make RNA to use same BLI functions,
so leaving that behind for now.
This is really close to BLI_mempool but uses an array to keep track of the
chunks of memory. There is no tagging necessary to clear the whole
structure so reuse is fast.
Naturally supports iteration but does not support freeing.
matrix inversion was changed in rB01c75c3765eb from own code to EIGEN
for performance reasons. EIGEN would return a zero matrix on failure
(resulting in the pivot always being at the object origin).
This brings back the "old" matrix inversion code (which has the benifit
of providing a partial solution which makes the local transform center
appear correct)
Reviewers: campbellbarton
Maniphest Tasks: T57767
Differential Revision: https://developer.blender.org/D4804
Unlike location and rotation, there is a meaningful definition of
overall/average scaling via the total change in the volume. This
adds an option to retrieve that via a single driver variable,
instead of having to use three and an expression.
Using the determinant to compute the volume scaling also allows
detecting flipping due to negative scale - this is impossible
to do via the three variable approach.
The volume_scale functions are added purely for code readability:
'volume scale factor' is easier to understand than determinant.
Differential Revision: https://developer.blender.org/D4803
