2023-08-16 00:20:26 +10:00
|
|
|
/* SPDX-FileCopyrightText: 2018 Blender Authors
|
2023-05-31 16:19:06 +02:00
|
|
|
*
|
|
|
|
|
* SPDX-License-Identifier: GPL-2.0-or-later */
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
|
2019-02-18 08:08:12 +11:00
|
|
|
/** \file
|
|
|
|
|
* \ingroup bke
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
*/
|
|
|
|
|
|
2023-08-02 22:14:18 +02:00
|
|
|
#include "BKE_subdiv.hh"
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
|
2018-08-01 15:43:57 +02:00
|
|
|
#include "DNA_mesh_types.h"
|
2018-08-14 12:17:10 +02:00
|
|
|
#include "DNA_modifier_types.h"
|
2018-08-01 15:43:57 +02:00
|
|
|
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
#include "BLI_utildefines.h"
|
|
|
|
|
|
2023-08-02 22:14:18 +02:00
|
|
|
#include "BKE_subdiv_modifier.hh"
|
OpenSubDiv: add support for an OpenGL evaluator
This evaluator is used in order to evaluate subdivision at render time, allowing for
faster renders of meshes with a subdivision surface modifier placed at the last
position in the modifier list.
When evaluating the subsurf modifier, we detect whether we can delegate evaluation
to the draw code. If so, the subdivision is first evaluated on the GPU using our own
custom evaluator (only the coarse data needs to be initially sent to the GPU), then,
buffers for the final `MeshBufferCache` are filled on the GPU using a set of
compute shaders. However, some buffers are still filled on the CPU side, if doing so
on the GPU is impractical (e.g. the line adjacency buffer used for x-ray, whose
logic is hardly GPU compatible).
This is done at the mesh buffer extraction level so that the result can be readily used
in the various OpenGL engines, without having to write custom geometry or tesselation
shaders.
We use our own subdivision evaluation shaders, instead of OpenSubDiv's vanilla one, in
order to control the data layout, and interpolation. For example, we store vertex colors
as compressed 16-bit integers, while OpenSubDiv's default evaluator only work for float
types.
In order to still access the modified geometry on the CPU side, for use in modifiers
or transform operators, a dedicated wrapper type is added `MESH_WRAPPER_TYPE_SUBD`.
Subdivision will be lazily evaluated via `BKE_object_get_evaluated_mesh` which will
create such a wrapper if possible. If the final subdivision surface is not needed on
the CPU side, `BKE_object_get_evaluated_mesh_no_subsurf` should be used.
Enabling or disabling GPU subdivision can be done through the user preferences (under
Viewport -> Subdivision).
See patch description for benchmarks.
Reviewed By: campbellbarton, jbakker, fclem, brecht, #eevee_viewport
Differential Revision: https://developer.blender.org/D12406
2021-12-27 16:34:47 +01:00
|
|
|
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
#include "MEM_guardedalloc.h"
|
|
|
|
|
|
2023-08-02 22:14:18 +02:00
|
|
|
#include "subdiv_converter.hh"
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
|
2023-12-10 20:18:40 -05:00
|
|
|
#include "opensubdiv_capi.hh"
|
|
|
|
|
#include "opensubdiv_converter_capi.hh"
|
|
|
|
|
#include "opensubdiv_evaluator_capi.hh"
|
2024-09-29 08:47:37 +02:00
|
|
|
#ifdef WITH_OPENSUBDIV
|
2024-09-27 19:01:12 +02:00
|
|
|
# include "opensubdiv_topology_refiner_capi.hh"
|
|
|
|
|
#endif
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
namespace blender::bke::subdiv {
|
|
|
|
|
|
2022-05-12 09:43:19 +02:00
|
|
|
/* --------------------------------------------------------------------
|
|
|
|
|
* Module.
|
|
|
|
|
*/
|
2020-05-18 16:42:59 +02:00
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
void init()
|
2020-05-18 16:42:59 +02:00
|
|
|
{
|
|
|
|
|
openSubdiv_init();
|
|
|
|
|
}
|
|
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
void exit()
|
2020-05-18 16:42:59 +02:00
|
|
|
{
|
|
|
|
|
openSubdiv_cleanup();
|
|
|
|
|
}
|
|
|
|
|
|
2022-05-12 09:43:19 +02:00
|
|
|
/* --------------------------------------------------------------------
|
|
|
|
|
* Conversion helpers.
|
|
|
|
|
*/
|
2019-01-14 11:11:42 +01:00
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
FVarLinearInterpolation fvar_interpolation_from_uv_smooth(int uv_smooth)
|
2018-08-14 12:17:10 +02:00
|
|
|
{
|
|
|
|
|
switch (uv_smooth) {
|
|
|
|
|
case SUBSURF_UV_SMOOTH_NONE:
|
|
|
|
|
return SUBDIV_FVAR_LINEAR_INTERPOLATION_ALL;
|
|
|
|
|
case SUBSURF_UV_SMOOTH_PRESERVE_CORNERS:
|
|
|
|
|
return SUBDIV_FVAR_LINEAR_INTERPOLATION_CORNERS_ONLY;
|
|
|
|
|
case SUBSURF_UV_SMOOTH_PRESERVE_CORNERS_AND_JUNCTIONS:
|
|
|
|
|
return SUBDIV_FVAR_LINEAR_INTERPOLATION_CORNERS_AND_JUNCTIONS;
|
|
|
|
|
case SUBSURF_UV_SMOOTH_PRESERVE_CORNERS_JUNCTIONS_AND_CONCAVE:
|
|
|
|
|
return SUBDIV_FVAR_LINEAR_INTERPOLATION_CORNERS_JUNCTIONS_AND_CONCAVE;
|
|
|
|
|
case SUBSURF_UV_SMOOTH_PRESERVE_BOUNDARIES:
|
|
|
|
|
return SUBDIV_FVAR_LINEAR_INTERPOLATION_BOUNDARIES;
|
|
|
|
|
case SUBSURF_UV_SMOOTH_ALL:
|
|
|
|
|
return SUBDIV_FVAR_LINEAR_INTERPOLATION_NONE;
|
|
|
|
|
}
|
2021-07-15 18:23:28 +10:00
|
|
|
BLI_assert_msg(0, "Unknown uv smooth flag");
|
2018-09-28 10:13:18 +02:00
|
|
|
return SUBDIV_FVAR_LINEAR_INTERPOLATION_ALL;
|
2018-08-14 12:17:10 +02:00
|
|
|
}
|
|
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
VtxBoundaryInterpolation vtx_boundary_interpolation_from_subsurf(int boundary_smooth)
|
2020-09-30 15:56:37 +02:00
|
|
|
{
|
|
|
|
|
switch (boundary_smooth) {
|
|
|
|
|
case SUBSURF_BOUNDARY_SMOOTH_PRESERVE_CORNERS:
|
|
|
|
|
return SUBDIV_VTX_BOUNDARY_EDGE_AND_CORNER;
|
|
|
|
|
case SUBSURF_BOUNDARY_SMOOTH_ALL:
|
|
|
|
|
return SUBDIV_VTX_BOUNDARY_EDGE_ONLY;
|
|
|
|
|
}
|
2021-07-15 18:23:28 +10:00
|
|
|
BLI_assert_msg(0, "Unknown boundary smooth flag");
|
2020-09-30 15:56:37 +02:00
|
|
|
return SUBDIV_VTX_BOUNDARY_EDGE_ONLY;
|
|
|
|
|
}
|
|
|
|
|
|
2022-05-12 09:43:19 +02:00
|
|
|
/* --------------------------------------------------------------------
|
|
|
|
|
* Settings.
|
|
|
|
|
*/
|
2019-01-15 15:34:11 +01:00
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
bool settings_equal(const Settings *settings_a, const Settings *settings_b)
|
2019-01-15 15:34:11 +01:00
|
|
|
{
|
|
|
|
|
return (settings_a->is_simple == settings_b->is_simple &&
|
|
|
|
|
settings_a->is_adaptive == settings_b->is_adaptive &&
|
|
|
|
|
settings_a->level == settings_b->level &&
|
|
|
|
|
settings_a->vtx_boundary_interpolation == settings_b->vtx_boundary_interpolation &&
|
|
|
|
|
settings_a->fvar_linear_interpolation == settings_b->fvar_linear_interpolation);
|
|
|
|
|
}
|
|
|
|
|
|
2022-05-12 09:43:19 +02:00
|
|
|
/* --------------------------------------------------------------------
|
|
|
|
|
* Construction.
|
|
|
|
|
*/
|
2019-01-14 11:11:42 +01:00
|
|
|
|
2019-01-15 15:34:11 +01:00
|
|
|
/* Creation from scratch. */
|
|
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
Subdiv *new_from_converter(const Settings *settings, OpenSubdiv_Converter *converter)
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
{
|
2024-09-27 19:01:12 +02:00
|
|
|
#ifdef WITH_OPENSUBDIV
|
2018-07-19 16:27:18 +02:00
|
|
|
SubdivStats stats;
|
2024-04-18 17:27:10 -04:00
|
|
|
stats_init(&stats);
|
|
|
|
|
stats_begin(&stats, SUBDIV_STATS_TOPOLOGY_REFINER_CREATION_TIME);
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
OpenSubdiv_TopologyRefinerSettings topology_refiner_settings;
|
|
|
|
|
topology_refiner_settings.level = settings->level;
|
|
|
|
|
topology_refiner_settings.is_adaptive = settings->is_adaptive;
|
2024-09-27 19:01:12 +02:00
|
|
|
blender::opensubdiv::TopologyRefinerImpl *osd_topology_refiner = nullptr;
|
2018-08-01 15:43:57 +02:00
|
|
|
if (converter->getNumVertices(converter) != 0) {
|
2024-09-27 19:01:12 +02:00
|
|
|
osd_topology_refiner = blender::opensubdiv::TopologyRefinerImpl::createFromConverter(
|
|
|
|
|
converter, topology_refiner_settings);
|
2018-08-01 15:43:57 +02:00
|
|
|
}
|
|
|
|
|
else {
|
|
|
|
|
/* TODO(sergey): Check whether original geometry had any vertices.
|
|
|
|
|
* The thing here is: OpenSubdiv can only deal with faces, but our
|
2019-01-14 11:11:42 +01:00
|
|
|
* side of subdiv also deals with loose vertices and edges. */
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
}
|
2023-01-19 21:40:50 -06:00
|
|
|
Subdiv *subdiv = MEM_cnew<Subdiv>(__func__);
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
subdiv->settings = *settings;
|
|
|
|
|
subdiv->topology_refiner = osd_topology_refiner;
|
2023-01-19 21:40:50 -06:00
|
|
|
subdiv->evaluator = nullptr;
|
|
|
|
|
subdiv->displacement_evaluator = nullptr;
|
2024-04-18 17:27:10 -04:00
|
|
|
stats_end(&stats, SUBDIV_STATS_TOPOLOGY_REFINER_CREATION_TIME);
|
2018-07-19 16:27:18 +02:00
|
|
|
subdiv->stats = stats;
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
return subdiv;
|
2024-09-27 19:01:12 +02:00
|
|
|
#else
|
|
|
|
|
UNUSED_VARS(settings, converter);
|
|
|
|
|
return nullptr;
|
|
|
|
|
#endif
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
}
|
|
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
Subdiv *new_from_mesh(const Settings *settings, const Mesh *mesh)
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
{
|
2023-12-20 02:21:48 +01:00
|
|
|
if (mesh->verts_num == 0) {
|
2023-01-19 21:40:50 -06:00
|
|
|
return nullptr;
|
2018-08-01 15:43:57 +02:00
|
|
|
}
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
OpenSubdiv_Converter converter;
|
2024-04-18 17:27:10 -04:00
|
|
|
converter_init_for_mesh(&converter, settings, mesh);
|
|
|
|
|
Subdiv *subdiv = new_from_converter(settings, &converter);
|
|
|
|
|
converter_free(&converter);
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
return subdiv;
|
|
|
|
|
}
|
|
|
|
|
|
2019-01-15 15:34:11 +01:00
|
|
|
/* Creation with cached-aware semantic. */
|
|
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
Subdiv *update_from_converter(Subdiv *subdiv,
|
|
|
|
|
const Settings *settings,
|
|
|
|
|
OpenSubdiv_Converter *converter)
|
2019-01-15 15:34:11 +01:00
|
|
|
{
|
2024-09-27 19:01:12 +02:00
|
|
|
#ifdef WITH_OPENSUBDIV
|
2019-01-15 15:34:11 +01:00
|
|
|
/* Check if the existing descriptor can be re-used. */
|
|
|
|
|
bool can_reuse_subdiv = true;
|
2023-01-19 21:40:50 -06:00
|
|
|
if (subdiv != nullptr && subdiv->topology_refiner != nullptr) {
|
2024-04-18 17:27:10 -04:00
|
|
|
if (!settings_equal(&subdiv->settings, settings)) {
|
2019-01-15 15:34:11 +01:00
|
|
|
can_reuse_subdiv = false;
|
|
|
|
|
}
|
|
|
|
|
else {
|
2024-04-18 17:27:10 -04:00
|
|
|
stats_begin(&subdiv->stats, SUBDIV_STATS_TOPOLOGY_COMPARE);
|
2024-09-27 19:01:12 +02:00
|
|
|
can_reuse_subdiv = subdiv->topology_refiner->isEqualToConverter(converter);
|
2024-04-18 17:27:10 -04:00
|
|
|
stats_end(&subdiv->stats, SUBDIV_STATS_TOPOLOGY_COMPARE);
|
2019-04-17 06:17:24 +02:00
|
|
|
}
|
2019-01-15 15:34:11 +01:00
|
|
|
}
|
|
|
|
|
else {
|
|
|
|
|
can_reuse_subdiv = false;
|
|
|
|
|
}
|
|
|
|
|
if (can_reuse_subdiv) {
|
|
|
|
|
return subdiv;
|
|
|
|
|
}
|
|
|
|
|
/* Create new subdiv. */
|
2023-01-19 21:40:50 -06:00
|
|
|
if (subdiv != nullptr) {
|
2024-04-18 17:27:10 -04:00
|
|
|
free(subdiv);
|
2019-01-15 15:34:11 +01:00
|
|
|
}
|
2024-04-18 17:27:10 -04:00
|
|
|
return new_from_converter(settings, converter);
|
2024-09-27 19:01:12 +02:00
|
|
|
#else
|
|
|
|
|
UNUSED_VARS(subdiv, settings, converter);
|
|
|
|
|
return nullptr;
|
|
|
|
|
#endif
|
2019-01-15 15:34:11 +01:00
|
|
|
}
|
|
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
Subdiv *update_from_mesh(Subdiv *subdiv, const Settings *settings, const Mesh *mesh)
|
2019-01-15 15:34:11 +01:00
|
|
|
{
|
|
|
|
|
OpenSubdiv_Converter converter;
|
2024-04-18 17:27:10 -04:00
|
|
|
converter_init_for_mesh(&converter, settings, mesh);
|
|
|
|
|
subdiv = update_from_converter(subdiv, settings, &converter);
|
|
|
|
|
converter_free(&converter);
|
2019-01-15 15:34:11 +01:00
|
|
|
return subdiv;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Memory release. */
|
|
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
void free(Subdiv *subdiv)
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
{
|
2024-09-27 19:01:12 +02:00
|
|
|
#ifdef WITH_OPENSUBDIV
|
2023-01-19 21:40:50 -06:00
|
|
|
if (subdiv->evaluator != nullptr) {
|
OpenSubDiv: add support for an OpenGL evaluator
This evaluator is used in order to evaluate subdivision at render time, allowing for
faster renders of meshes with a subdivision surface modifier placed at the last
position in the modifier list.
When evaluating the subsurf modifier, we detect whether we can delegate evaluation
to the draw code. If so, the subdivision is first evaluated on the GPU using our own
custom evaluator (only the coarse data needs to be initially sent to the GPU), then,
buffers for the final `MeshBufferCache` are filled on the GPU using a set of
compute shaders. However, some buffers are still filled on the CPU side, if doing so
on the GPU is impractical (e.g. the line adjacency buffer used for x-ray, whose
logic is hardly GPU compatible).
This is done at the mesh buffer extraction level so that the result can be readily used
in the various OpenGL engines, without having to write custom geometry or tesselation
shaders.
We use our own subdivision evaluation shaders, instead of OpenSubDiv's vanilla one, in
order to control the data layout, and interpolation. For example, we store vertex colors
as compressed 16-bit integers, while OpenSubDiv's default evaluator only work for float
types.
In order to still access the modified geometry on the CPU side, for use in modifiers
or transform operators, a dedicated wrapper type is added `MESH_WRAPPER_TYPE_SUBD`.
Subdivision will be lazily evaluated via `BKE_object_get_evaluated_mesh` which will
create such a wrapper if possible. If the final subdivision surface is not needed on
the CPU side, `BKE_object_get_evaluated_mesh_no_subsurf` should be used.
Enabling or disabling GPU subdivision can be done through the user preferences (under
Viewport -> Subdivision).
See patch description for benchmarks.
Reviewed By: campbellbarton, jbakker, fclem, brecht, #eevee_viewport
Differential Revision: https://developer.blender.org/D12406
2021-12-27 16:34:47 +01:00
|
|
|
const eOpenSubdivEvaluator evaluator_type = subdiv->evaluator->type;
|
|
|
|
|
if (evaluator_type != OPENSUBDIV_EVALUATOR_CPU) {
|
|
|
|
|
/* Let the draw code do the freeing, to ensure that the OpenGL context is valid. */
|
|
|
|
|
BKE_subsurf_modifier_free_gpu_cache_cb(subdiv);
|
|
|
|
|
return;
|
|
|
|
|
}
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
openSubdiv_deleteEvaluator(subdiv->evaluator);
|
|
|
|
|
}
|
2024-09-27 19:01:12 +02:00
|
|
|
delete subdiv->topology_refiner;
|
2024-04-18 17:27:10 -04:00
|
|
|
displacement_detach(subdiv);
|
2023-01-19 21:40:50 -06:00
|
|
|
if (subdiv->cache_.face_ptex_offset != nullptr) {
|
2018-09-06 15:39:27 +02:00
|
|
|
MEM_freeN(subdiv->cache_.face_ptex_offset);
|
|
|
|
|
}
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
MEM_freeN(subdiv);
|
2024-09-27 19:01:12 +02:00
|
|
|
#else
|
|
|
|
|
UNUSED_VARS(subdiv);
|
|
|
|
|
#endif
|
Subsurf: Begin new subdivision surface module
The idea is to use this as a replacement of old CCG, now it is
based on OpenSubdiv. The goal is to reduce any possible overhead
which was happening with OpenSubdiv used by CCG.
Currently implemented/supported:
- Creation from mesh, including topology on OpenSubdiv side,
its refinement.
- Evaluation of limit point, first order derivatives, normal,
and face-varying data for individual coarse position.
- Evaluation of whole patches.
Currently not optimized, uses evaluation of individual coarse
positions.
- Creation of Mesh from subdiv, with all geometry being real:
all mvert, medge, mloop, and mpoly.
This includes custom data interpolation, but all faces currently
are getting separated (they are converted to ptex patches, which
we need to weld back together).
Still need to support lighter weights grids and such, but this
is already a required part to have subsurf working in the middle
of modifier stack.
Annoying part is ifdef all over the place, to keep it compilable
when OpenSubdiv is disabled. More cleaner approach would be to
have stub API for OpenSubdiv, so everything gets ifdef-ed in a
much fewer places.
2018-07-17 18:07:26 +02:00
|
|
|
}
|
2018-09-06 15:39:27 +02:00
|
|
|
|
2022-05-12 09:43:19 +02:00
|
|
|
/* --------------------------------------------------------------------
|
|
|
|
|
* Topology helpers.
|
|
|
|
|
*/
|
2019-01-14 11:11:42 +01:00
|
|
|
|
2024-04-18 17:27:10 -04:00
|
|
|
int *face_ptex_offset_get(Subdiv *subdiv)
|
2018-09-06 15:39:27 +02:00
|
|
|
{
|
2024-04-18 22:27:35 -04:00
|
|
|
#ifdef WITH_OPENSUBDIV
|
2023-01-19 21:40:50 -06:00
|
|
|
if (subdiv->cache_.face_ptex_offset != nullptr) {
|
2018-09-06 15:39:27 +02:00
|
|
|
return subdiv->cache_.face_ptex_offset;
|
|
|
|
|
}
|
2024-09-27 19:01:12 +02:00
|
|
|
const blender::opensubdiv::TopologyRefinerImpl *topology_refiner = subdiv->topology_refiner;
|
2023-01-19 21:40:50 -06:00
|
|
|
if (topology_refiner == nullptr) {
|
|
|
|
|
return nullptr;
|
2018-09-21 12:18:50 +02:00
|
|
|
}
|
2024-09-27 19:01:12 +02:00
|
|
|
const int num_coarse_faces = topology_refiner->base_level().GetNumFaces();
|
2023-01-19 21:40:50 -06:00
|
|
|
subdiv->cache_.face_ptex_offset = static_cast<int *>(
|
|
|
|
|
MEM_malloc_arrayN(num_coarse_faces + 1, sizeof(int), __func__));
|
2018-09-06 15:39:27 +02:00
|
|
|
int ptex_offset = 0;
|
|
|
|
|
for (int face_index = 0; face_index < num_coarse_faces; face_index++) {
|
2024-09-27 19:01:12 +02:00
|
|
|
const int face_size = topology_refiner->base_level().GetFaceVertices(face_index).size();
|
|
|
|
|
const int num_ptex_faces = face_size == 4 ? 1 : face_size;
|
2018-09-06 15:39:27 +02:00
|
|
|
subdiv->cache_.face_ptex_offset[face_index] = ptex_offset;
|
|
|
|
|
ptex_offset += num_ptex_faces;
|
|
|
|
|
}
|
OpenSubDiv: add support for an OpenGL evaluator
This evaluator is used in order to evaluate subdivision at render time, allowing for
faster renders of meshes with a subdivision surface modifier placed at the last
position in the modifier list.
When evaluating the subsurf modifier, we detect whether we can delegate evaluation
to the draw code. If so, the subdivision is first evaluated on the GPU using our own
custom evaluator (only the coarse data needs to be initially sent to the GPU), then,
buffers for the final `MeshBufferCache` are filled on the GPU using a set of
compute shaders. However, some buffers are still filled on the CPU side, if doing so
on the GPU is impractical (e.g. the line adjacency buffer used for x-ray, whose
logic is hardly GPU compatible).
This is done at the mesh buffer extraction level so that the result can be readily used
in the various OpenGL engines, without having to write custom geometry or tesselation
shaders.
We use our own subdivision evaluation shaders, instead of OpenSubDiv's vanilla one, in
order to control the data layout, and interpolation. For example, we store vertex colors
as compressed 16-bit integers, while OpenSubDiv's default evaluator only work for float
types.
In order to still access the modified geometry on the CPU side, for use in modifiers
or transform operators, a dedicated wrapper type is added `MESH_WRAPPER_TYPE_SUBD`.
Subdivision will be lazily evaluated via `BKE_object_get_evaluated_mesh` which will
create such a wrapper if possible. If the final subdivision surface is not needed on
the CPU side, `BKE_object_get_evaluated_mesh_no_subsurf` should be used.
Enabling or disabling GPU subdivision can be done through the user preferences (under
Viewport -> Subdivision).
See patch description for benchmarks.
Reviewed By: campbellbarton, jbakker, fclem, brecht, #eevee_viewport
Differential Revision: https://developer.blender.org/D12406
2021-12-27 16:34:47 +01:00
|
|
|
subdiv->cache_.face_ptex_offset[num_coarse_faces] = ptex_offset;
|
2018-09-06 15:39:27 +02:00
|
|
|
return subdiv->cache_.face_ptex_offset;
|
2024-04-18 22:27:35 -04:00
|
|
|
#else
|
|
|
|
|
UNUSED_VARS(subdiv);
|
|
|
|
|
return nullptr;
|
|
|
|
|
#endif
|
2018-09-06 15:39:27 +02:00
|
|
|
}
|
2024-04-18 17:27:10 -04:00
|
|
|
|
|
|
|
|
} // namespace blender::bke::subdiv
|