Blender Git Loki

Kaikki Blender Git kommitit.

Page: 4609 / 8462

November 25, 2015, 10:54 (GMT)
PyAPI: remove workaround for Py3.4 bug
November 25, 2015, 10:51 (GMT)
MOD_weightvg: switch from OMP to BLI_task's parallelized 'for loop'.

Also gives about 5-6% speedup...
November 25, 2015, 10:43 (GMT)
Merge branch 'master' into GPencil_Editing_Stage3
November 25, 2015, 10:27 (GMT)
Added an 'entry point' instruction number to functions.

This is currently always 0, but will be used in the future to allow
storing per-element functions along with the main code. These can then
be called within kernels.
November 25, 2015, 10:12 (GMT)
Enabled generic math nodes for modifier node trees.
November 25, 2015, 10:01 (GMT)
BLI_task: BLI_task_parallel_range_ex: add some per-chunk userdata.

This mimics OpenMP's 'firstprivate' feature. It is sometimes handy to have some persistent local data during a whole chunk.

Reviewers: sergey

Reviewed By: sergey

Subscribers: campbellbarton

Differential Revision: https://developer.blender.org/D1635
November 25, 2015, 09:51 (GMT)
Fix for static empty mesh variable, used as default value of meshes on the stack.

This creates a (minor) mem leak when statically creating the mesh. Instead it
should be constructed in a init/free pair of functions.

Because this variable is used in all different compile units, each of them has
to create their own global 'empty_mesh' variable.
November 25, 2015, 09:01 (GMT)
Fixed missing delete call for unused nodes.
November 25, 2015, 08:58 (GMT)
Free the modifier bvm function after use.

Note that eventually these should be cached to avoid unnecessary
recompilation, but for now they get compiled on every modifier eval.
November 25, 2015, 08:53 (GMT)
Use the guardedalloc new/delete overloads for relevant classes.
November 25, 2015, 08:46 (GMT)
Fix T46782: Updating Shaders very slow with complex nodegraph

The issue was caused by not really optimal graph traversal for gathering nodes
dependencies which could have exponential complexity with a long tree branches
connected with multiple connections between them.

Now we optimize the depth traversal and perform early output if the node was
already traversed.

Please note that this adds some limitations to the use of SVM compiler's
find_dependencies() in the cases when skip_node is not NULL and one wants to
perform dependencies find sequentially with the same set. This doesn't happen
in the code, but one should be aware of this.
November 25, 2015, 08:29 (GMT)
Fixed overzealous assert for input stack assignments in codegen.

For constant inputs no stack index is needed. Previous to node pruning
this did not trigger the assert because there were still pass nodes
allocating the stack indices.
November 25, 2015, 08:28 (GMT)
Node tree pruning function for removing unused nodes.

This also removes the placeholder 'pass' nodes, which were already
being skipped and made unused.
November 25, 2015, 08:09 (GMT)
Changed node instance storage in the node graph to pointers.

This makes it much easier to optimize the graph by removing and reconnecting nodes.
November 25, 2015, 08:07 (GMT)
Cycles: Ensure order of shader nodes in the dependnecies set

The issue was than nodes dependencies were stored as set<ShaderNode*> which
is actually a so called "strict weak ordered", meaning order of nodes in
the set is strictly defined, but based on the ShaderNode pointer. This means
that between different render invokations order of original nodes could be
different due to different pointers allocated for ShaderNode.

This commit makes it so dependencies and maps used for ShaderNodes are based
on the node->id which has much more predictable order. It's still possible
to trick the system by doing some crazy edits during viewport rendfer and
cause difference between viewport and final render stacks.

Reviewers: brecht

Reviewed By: brecht

Subscribers: LazyDodo

Differential Revision: https://developer.blender.org/D1630
November 25, 2015, 08:01 (GMT)
Cycles: Fix wrong volume stack update with SSS object intersecting the volume

There's no need in moving ray at all, stack should always be updated from the
original hit point to the scattered one.
November 25, 2015, 08:01 (GMT)
Cycles: Trace indirect subsurface rays by restarting the integrator loop

This gives much lower stack usage on GPU and reduces kernel memory size to
around 448MB on GTX560Ti (comparing to 652MB with previous commit and 946MB
with official release). There's also a barely measurable speedup of around
5%, but this is to be confirmed still.

At this stage we're using only ~3% for the experimental kernel and SSS
rendering seems to be faster by 40% and after some further testing we might
consider making SSS and CMJ official features and remove experimental
precompiled kernels.
November 25, 2015, 08:01 (GMT)
Cycles: Delay shooting SSS indirect rays

The idea is to delay shooting indirect rays for the SSS sampling and
trace them after the main integration loop was finished.

This reduces GPU stack usage even further and brings it down to around
652MB (comparing to 722MB before the change and 946MB with previous
stable release).

This also solves the speed regression happened in the previous commit
and now simple SSS scene (SSS suzanne on the floor) renders in 0:50
(comparing to 1:16 with previous commit and 1:03 with official release).
November 25, 2015, 08:01 (GMT)
Cysles: Avoid having ShaderData on the stack

This commit introduces a SSS-oriented intersection structure which is replacing
old logic of having separate arrays for just intersections and shader data and
encapsulates all the data needed for SSS evaluation.

This giver a huge stack memory saving on GPU. In own experiments it gave 25%
memory usage reduction on GTX560Ti (722MB vs. 946MB).

Unfortunately, this gave some performance loss of 20% which only happens on GPU.
This is perhaps due to different memory access pattern. Will be solved in the
future, hopefully.

Famous saying: won in memory - lost in time (which is also valid in other way
around).
Revision e6fff42 by Mike Erwin (master)
November 25, 2015, 06:49 (GMT)
OpenGL: set geometry shader input length implicitly

Input array length is implicitly set at link time, based on the geometry
shader's layout. Specifying the wrong value here is an error; specifying
no value is the same as getting it right. (inspired by a recent codegen
change)
Tehnyt: Miika HämäläinenViimeksi päivitetty: 07.11.2014 14:18MiikaH:n Sivut a.k.a. MiikaHweb | 2003-2021