Revision f0e49ff by Bastien Montagne May 9, 2016, 13:01 (GMT) |
Revert "Squashed commit of the atomic-operations branch." This reverts commit af46348b8207df032be67f6434af6ce20286bc50. |
Revision af46348 by Bastien Montagne May 9, 2016, 13:01 (GMT) |
Squashed commit of the atomic-operations branch. |
May 9, 2016, 12:54 (GMT) |
Task scheduler: Don't calloc in performance-critical areas Majority of the fields are being overwritten anyway, so calloc it kinda waste of CPU ticks. |
Revision 86e2e2e by Bastien Montagne May 9, 2016, 12:29 (GMT) |
Revert "Squashed commit of the atomic-operations branch." This reverts commit 79cbfbf9ee0354f8b3a30d64e5a47f3d083be618. |
Revision 79cbfbf by Bastien Montagne May 9, 2016, 12:28 (GMT) |
Squashed commit of the atomic-operations branch. |
Revision a801335 by Bastien Montagne May 9, 2016, 12:28 (GMT) |
Merge branch 'master' into experimental-build |
May 9, 2016, 10:42 (GMT) |
Depsgraph: Store node input/output links in a vector rather than in set Set is much slower to iterate through (due to cache misses and such) and the only advantage of using set is faster removal of link. However, we are iterating links much much more often than removing them, and even when we are removing links we don't really need to remove link from nodes which it connects -- we don't support partial depsgraph updates, so removing links from nodes on destruction is a waste of time. If we ever want to support partial updates we can have dedicated function to remove link from nodes it connects. This gives a surprising increase of fps from 42 to 56 with test file from Mr. J.P.Bouza (blenrig_for_debugging.blend). Surprising because old DEG is actually slower here (52 fps). Didn't see any regressions (and don't see why they will happen), so let's ask our riggers and animators to perform further speed tests ;) |
May 9, 2016, 09:58 (GMT) |
Depsgraph: Avoid having per-node lock when scheduling children Use atomic operations instead, should in theory improve timing of scheduling. However, probably not so visible yet because actual task scheduling still have some locks and memory allocations. Baby steps, what would i say. |
May 9, 2016, 09:45 (GMT) |
Atomics: Add atomic_fetch_and_or_uint8() function Hopefully it compiles on all platforms still.. |
May 9, 2016, 07:54 (GMT) |
Fix makesdna not checking alignment for a non-native platform This was causing alignment issues which were only visible on a platform of particular bitness, so it was easy to break stuff for 32bit when working on 64bit platform and vice versa. |
May 9, 2016, 07:42 (GMT) |
Fixes for GPencil Copy and Paste * Fix "Attempt to free NULL pointer" when copying strokes for the first time * Fix poll callback on "paste" operator, so that it is possible to paste strokes when there are no editable strokes visible. |
May 9, 2016, 07:16 (GMT) |
Cleanup: Avoid duplicative defines for CPU textures, use the ones from util_texture.h Also includes some further byte -> byte4 renaming, missed that in last commit. |
May 9, 2016, 07:12 (GMT) |
Moved type definition functions for LLVM types into own file. |
May 9, 2016, 06:21 (GMT) |
Fix error introduced by removing faces before executing bridge Commit 86abddc9, caused an error when the face-region included boundary edges. Since removing the faces first, caused the edges along the boundaries to be removed. Add support for deleting faces and internal edges, that keeps boundaries. |
May 9, 2016, 06:17 (GMT) |
Disable IR module loading for blenvm by default (with a compile option). Base modules for nodes will instead be declared as external functions with regular linkage through C code. This means some potential for low-level optimization might be lost, but this shouldn't be a big problem as long as node functions don't get too complex (which means they should be split up in the first place). IR module loading is not the real issue, but generating the IR code from C/C++ code in a generic way is not so easy. In the future we may actually generate most code for basic operators and types directly inside blenvm and only use precompiled modules for more complex functions. |
May 9, 2016, 05:46 (GMT) |
Fix error in recent snap refactor Crash in edge snapping |
May 9, 2016, 05:46 (GMT) |
Fix crash introduced w/ snap refactor Create snapping context when in the 3d view, transforming pose bones and grease-pencil crashed. |
May 9, 2016, 05:23 (GMT) |
GPencil: Added operators to select first and last points of strokes These are useful for removing overshoots at the end of closed strokes (for fills) |
May 9, 2016, 05:23 (GMT) |
Graph Editor: Ctrl-Click keyframing now deselects all existing keyframes by default Now, when creating new keyframes in the graph editor by ctrl-clicking, only the newly created keyframes will be selected. This is a little workflow tweak to make it faster to work, as you no longer have to deselect all, and then re-select the newly added keys in order to manipulate them. The old behaviour (not modifying the selection status of the old keys) has been kept, but is now available via Shift-Ctrl-Click. Feature request from @Shhlife |
May 9, 2016, 00:22 (GMT) |
Cleanup: More byte -> byte4 renaming for consistency. |
|
|
|


Master Commits
MiikaHweb | 2003-2021