Files
test2/source/blender/blenkernel/intern/pointcloud.cc

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

411 lines
13 KiB
C++
Raw Normal View History

/* SPDX-FileCopyrightText: 2023 Blender Authors
*
* SPDX-License-Identifier: GPL-2.0-or-later */
/** \file
* \ingroup bke
*/
#include "MEM_guardedalloc.h"
#include "DNA_defaults.h"
#include "DNA_material_types.h"
#include "DNA_object_types.h"
#include "DNA_pointcloud_types.h"
#include "BLI_bounds.hh"
#include "BLI_index_range.hh"
#include "BLI_listbase.h"
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
#include "BLI_math_vector.hh"
#include "BLI_rand.h"
#include "BLI_span.hh"
#include "BLI_string.h"
#include "BLI_task.hh"
#include "BLI_utildefines.h"
#include "BLI_vector.hh"
#include "BKE_anim_data.h"
#include "BKE_customdata.h"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
#include "BKE_geometry_set.hh"
#include "BKE_global.h"
#include "BKE_idtype.h"
#include "BKE_lib_id.h"
#include "BKE_lib_query.h"
#include "BKE_lib_remap.h"
#include "BKE_main.h"
#include "BKE_mesh_wrapper.hh"
#include "BKE_modifier.h"
#include "BKE_object.h"
#include "BKE_pointcloud.h"
#include "BLT_translation.h"
#include "DEG_depsgraph_query.h"
#include "BLO_read_write.hh"
using blender::float3;
using blender::IndexRange;
using blender::Span;
using blender::Vector;
/* PointCloud datablock */
static void pointcloud_random(PointCloud *pointcloud);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const char *POINTCLOUD_ATTR_POSITION = "position";
const char *POINTCLOUD_ATTR_RADIUS = "radius";
static void pointcloud_init_data(ID *id)
{
PointCloud *pointcloud = (PointCloud *)id;
BLI_assert(MEMCMP_STRUCT_AFTER_IS_ZERO(pointcloud, id));
MEMCPY_STRUCT_AFTER(pointcloud, DNA_struct_default_get(PointCloud), id);
CustomData_reset(&pointcloud->pdata);
pointcloud->attributes_for_write().add<float3>(
"position", ATTR_DOMAIN_POINT, blender::bke::AttributeInitConstruct());
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
pointcloud->runtime = new blender::bke::PointCloudRuntime();
}
Custom Data: support implicit sharing for custom data layers This integrates the new implicit-sharing system (from fbcddfcd68adc72f) with `CustomData`. Now the potentially long arrays referenced by custom data layers can be shared between different systems but most importantly between different geometries. This makes e.g. copying a mesh much cheaper because none of the attributes has to be copied. Only when an attribute is modified does it have to be copied. Also see the original design task: #95845. This reduces memory and improves performance by avoiding unnecessary data copies. For example, the used memory after loading a highly subdivided mesh is reduced from 2.4GB to 1.79GB. This is about 25% less which is the expected amount because in `main` there are 4 copies of the data: 1. The original data which is allocated when the file is loaded. 2. The copy for the depsgraph allocated during depsgraph evaluation. 3. The copy for the undo system allocated when the first undo step is created right after loading the file. 4. GPU buffers allocated for drawing. This patch only gets rid of copy number 2 for the depsgraph. In theory the other copies can be removed as part of follow up PRs as well though. ----- The patch has three main components: * Slightly modified `CustomData` API to make it work better with implicit sharing: * `CD_REFERENCE` and `CD_DUPLICATE` have been removed because they are meaningless when implicit-sharing is used. * `CD_ASSIGN` has been removed as well because it's not an allocation type anyway. The functionality of using existing arrays as custom data layers has not been removed though. * This can still be done with `CustomData_add_layer_with_data` which also has a new argument that allows passing in information about whether the array is shared. * `CD_FLAG_NOFREE` has been removed because it's no longer necessary. It only existed because of `CD_REFERENCE`. * `CustomData_copy` and `CustomData_merge` have been split up into a functions that do copy the actual attribute values and those that do not. The latter functions now have the `_layout` suffix (e.g. `CustomData_copy_layout`). * Changes in `customdata.cc` to make it actually use implicit-sharing. * Changes in various other files to adapt to the changes in `BKE_customdata.h`. Pull Request: https://projects.blender.org/blender/blender/pulls/106228
2023-04-13 14:57:57 +02:00
static void pointcloud_copy_data(Main * /*bmain*/,
ID *id_dst,
const ID *id_src,
const int /*flag*/)
{
PointCloud *pointcloud_dst = (PointCloud *)id_dst;
const PointCloud *pointcloud_src = (const PointCloud *)id_src;
pointcloud_dst->mat = static_cast<Material **>(MEM_dupallocN(pointcloud_src->mat));
Custom Data: support implicit sharing for custom data layers This integrates the new implicit-sharing system (from fbcddfcd68adc72f) with `CustomData`. Now the potentially long arrays referenced by custom data layers can be shared between different systems but most importantly between different geometries. This makes e.g. copying a mesh much cheaper because none of the attributes has to be copied. Only when an attribute is modified does it have to be copied. Also see the original design task: #95845. This reduces memory and improves performance by avoiding unnecessary data copies. For example, the used memory after loading a highly subdivided mesh is reduced from 2.4GB to 1.79GB. This is about 25% less which is the expected amount because in `main` there are 4 copies of the data: 1. The original data which is allocated when the file is loaded. 2. The copy for the depsgraph allocated during depsgraph evaluation. 3. The copy for the undo system allocated when the first undo step is created right after loading the file. 4. GPU buffers allocated for drawing. This patch only gets rid of copy number 2 for the depsgraph. In theory the other copies can be removed as part of follow up PRs as well though. ----- The patch has three main components: * Slightly modified `CustomData` API to make it work better with implicit sharing: * `CD_REFERENCE` and `CD_DUPLICATE` have been removed because they are meaningless when implicit-sharing is used. * `CD_ASSIGN` has been removed as well because it's not an allocation type anyway. The functionality of using existing arrays as custom data layers has not been removed though. * This can still be done with `CustomData_add_layer_with_data` which also has a new argument that allows passing in information about whether the array is shared. * `CD_FLAG_NOFREE` has been removed because it's no longer necessary. It only existed because of `CD_REFERENCE`. * `CustomData_copy` and `CustomData_merge` have been split up into a functions that do copy the actual attribute values and those that do not. The latter functions now have the `_layout` suffix (e.g. `CustomData_copy_layout`). * Changes in `customdata.cc` to make it actually use implicit-sharing. * Changes in various other files to adapt to the changes in `BKE_customdata.h`. Pull Request: https://projects.blender.org/blender/blender/pulls/106228
2023-04-13 14:57:57 +02:00
CustomData_copy(
&pointcloud_src->pdata, &pointcloud_dst->pdata, CD_MASK_ALL, pointcloud_dst->totpoint);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
pointcloud_dst->runtime = new blender::bke::PointCloudRuntime();
pointcloud_dst->runtime->bounds_cache = pointcloud_src->runtime->bounds_cache;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
pointcloud_dst->batch_cache = nullptr;
}
static void pointcloud_free_data(ID *id)
{
PointCloud *pointcloud = (PointCloud *)id;
BKE_animdata_free(&pointcloud->id, false);
BKE_pointcloud_batch_cache_free(pointcloud);
CustomData_free(&pointcloud->pdata, pointcloud->totpoint);
MEM_SAFE_FREE(pointcloud->mat);
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
delete pointcloud->runtime;
}
static void pointcloud_foreach_id(ID *id, LibraryForeachIDData *data)
{
PointCloud *pointcloud = (PointCloud *)id;
for (int i = 0; i < pointcloud->totcol; i++) {
BKE_LIB_FOREACHID_PROCESS_IDSUPER(data, pointcloud->mat[i], IDWALK_CB_USER);
}
}
static void pointcloud_blend_write(BlendWriter *writer, ID *id, const void *id_address)
{
PointCloud *pointcloud = (PointCloud *)id;
Vector<CustomDataLayer, 16> point_layers;
CustomData_blend_write_prepare(pointcloud->pdata, point_layers);
/* Write LibData */
BLO_write_id_struct(writer, PointCloud, id_address, &pointcloud->id);
BKE_id_blend_write(writer, &pointcloud->id);
/* Direct data */
CustomData_blend_write(writer,
&pointcloud->pdata,
point_layers,
pointcloud->totpoint,
CD_MASK_ALL,
&pointcloud->id);
BLO_write_pointer_array(writer, pointcloud->totcol, pointcloud->mat);
}
static void pointcloud_blend_read_data(BlendDataReader *reader, ID *id)
{
PointCloud *pointcloud = (PointCloud *)id;
/* Geometry */
CustomData_blend_read(reader, &pointcloud->pdata, pointcloud->totpoint);
/* Materials */
BLO_read_pointer_array(reader, (void **)&pointcloud->mat);
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
pointcloud->runtime = new blender::bke::PointCloudRuntime();
}
IDTypeInfo IDType_ID_PT = {
/*id_code*/ ID_PT,
/*id_filter*/ FILTER_ID_PT,
/*main_listbase_index*/ INDEX_ID_PT,
/*struct_size*/ sizeof(PointCloud),
/*name*/ "PointCloud",
/*name_plural*/ "pointclouds",
/*translation_context*/ BLT_I18NCONTEXT_ID_POINTCLOUD,
/*flags*/ IDTYPE_FLAGS_APPEND_IS_REUSABLE,
/*asset_type_info*/ nullptr,
/*init_data*/ pointcloud_init_data,
/*copy_data*/ pointcloud_copy_data,
/*free_data*/ pointcloud_free_data,
/*make_local*/ nullptr,
/*foreach_id*/ pointcloud_foreach_id,
/*foreach_cache*/ nullptr,
/*foreach_path*/ nullptr,
/*owner_pointer_get*/ nullptr,
/*blend_write*/ pointcloud_blend_write,
/*blend_read_data*/ pointcloud_blend_read_data,
/*blend_read_after_liblink*/ nullptr,
/*blend_read_undo_preserve*/ nullptr,
/*lib_override_apply_post*/ nullptr,
};
static void pointcloud_random(PointCloud *pointcloud)
{
BLI_assert(pointcloud->totpoint == 0);
pointcloud->totpoint = 400;
CustomData_realloc(&pointcloud->pdata, 0, pointcloud->totpoint);
RNG *rng = BLI_rng_new(0);
blender::bke::MutableAttributeAccessor attributes = pointcloud->attributes_for_write();
blender::MutableSpan<float3> positions = pointcloud->positions_for_write();
blender::bke::SpanAttributeWriter<float> radii =
attributes.lookup_or_add_for_write_only_span<float>(POINTCLOUD_ATTR_RADIUS,
ATTR_DOMAIN_POINT);
for (const int i : positions.index_range()) {
positions[i] = float3(BLI_rng_get_float(rng), BLI_rng_get_float(rng), BLI_rng_get_float(rng)) *
2.0f -
1.0f;
radii.span[i] = 0.05f * BLI_rng_get_float(rng);
}
radii.finish();
BLI_rng_free(rng);
}
void *BKE_pointcloud_add(Main *bmain, const char *name)
{
PointCloud *pointcloud = static_cast<PointCloud *>(BKE_id_new(bmain, ID_PT, name));
return pointcloud;
}
void *BKE_pointcloud_add_default(Main *bmain, const char *name)
{
PointCloud *pointcloud = static_cast<PointCloud *>(BKE_libblock_alloc(bmain, ID_PT, name, 0));
pointcloud_init_data(&pointcloud->id);
pointcloud_random(pointcloud);
return pointcloud;
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
PointCloud *BKE_pointcloud_new_nomain(const int totpoint)
{
PointCloud *pointcloud = static_cast<PointCloud *>(BKE_libblock_alloc(
nullptr, ID_PT, BKE_idtype_idcode_to_name(ID_PT), LIB_ID_CREATE_LOCALIZE));
pointcloud_init_data(&pointcloud->id);
CustomData_realloc(&pointcloud->pdata, 0, totpoint);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
pointcloud->totpoint = totpoint;
return pointcloud;
}
void BKE_pointcloud_nomain_to_pointcloud(PointCloud *pointcloud_src, PointCloud *pointcloud_dst)
{
BLI_assert(pointcloud_src->id.tag & LIB_TAG_NO_MAIN);
CustomData_free(&pointcloud_dst->pdata, pointcloud_dst->totpoint);
const int totpoint = pointcloud_dst->totpoint = pointcloud_src->totpoint;
Custom Data: support implicit sharing for custom data layers This integrates the new implicit-sharing system (from fbcddfcd68adc72f) with `CustomData`. Now the potentially long arrays referenced by custom data layers can be shared between different systems but most importantly between different geometries. This makes e.g. copying a mesh much cheaper because none of the attributes has to be copied. Only when an attribute is modified does it have to be copied. Also see the original design task: #95845. This reduces memory and improves performance by avoiding unnecessary data copies. For example, the used memory after loading a highly subdivided mesh is reduced from 2.4GB to 1.79GB. This is about 25% less which is the expected amount because in `main` there are 4 copies of the data: 1. The original data which is allocated when the file is loaded. 2. The copy for the depsgraph allocated during depsgraph evaluation. 3. The copy for the undo system allocated when the first undo step is created right after loading the file. 4. GPU buffers allocated for drawing. This patch only gets rid of copy number 2 for the depsgraph. In theory the other copies can be removed as part of follow up PRs as well though. ----- The patch has three main components: * Slightly modified `CustomData` API to make it work better with implicit sharing: * `CD_REFERENCE` and `CD_DUPLICATE` have been removed because they are meaningless when implicit-sharing is used. * `CD_ASSIGN` has been removed as well because it's not an allocation type anyway. The functionality of using existing arrays as custom data layers has not been removed though. * This can still be done with `CustomData_add_layer_with_data` which also has a new argument that allows passing in information about whether the array is shared. * `CD_FLAG_NOFREE` has been removed because it's no longer necessary. It only existed because of `CD_REFERENCE`. * `CustomData_copy` and `CustomData_merge` have been split up into a functions that do copy the actual attribute values and those that do not. The latter functions now have the `_layout` suffix (e.g. `CustomData_copy_layout`). * Changes in `customdata.cc` to make it actually use implicit-sharing. * Changes in various other files to adapt to the changes in `BKE_customdata.h`. Pull Request: https://projects.blender.org/blender/blender/pulls/106228
2023-04-13 14:57:57 +02:00
CustomData_copy(&pointcloud_src->pdata, &pointcloud_dst->pdata, CD_MASK_ALL, totpoint);
BKE_id_free(nullptr, pointcloud_src);
}
std::optional<blender::Bounds<blender::float3>> PointCloud::bounds_min_max() const
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
using namespace blender;
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
using namespace blender::bke;
if (this->totpoint == 0) {
return std::nullopt;
}
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
this->runtime->bounds_cache.ensure([&](Bounds<float3> &r_bounds) {
const AttributeAccessor attributes = this->attributes();
const Span<float3> positions = this->positions();
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
if (attributes.contains(POINTCLOUD_ATTR_RADIUS)) {
const VArraySpan radii = *attributes.lookup<float>(POINTCLOUD_ATTR_RADIUS);
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
r_bounds = *bounds::min_max_with_radii(positions, radii);
}
else {
r_bounds = *bounds::min_max(positions);
}
});
return this->runtime->bounds_cache.data();
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
BoundBox *BKE_pointcloud_boundbox_get(Object *ob)
{
using namespace blender;
BLI_assert(ob->type == OB_POINTCLOUD);
if (ob->runtime.bb != nullptr && (ob->runtime.bb->flag & BOUNDBOX_DIRTY) == 0) {
return ob->runtime.bb;
}
if (ob->runtime.bb == nullptr) {
ob->runtime.bb = MEM_cnew<BoundBox>(__func__);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
std::optional<Bounds<float3>> bounds;
if (ob->runtime.geometry_set_eval) {
bounds = ob->runtime.geometry_set_eval->compute_boundbox_without_instances();
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
else {
const PointCloud *pointcloud = static_cast<PointCloud *>(ob->data);
bounds = pointcloud->bounds_min_max();
}
if (bounds) {
BKE_boundbox_init_from_minmax(ob->runtime.bb, bounds->min, bounds->max);
}
else {
BKE_boundbox_init_from_minmax(ob->runtime.bb, float3(-1), float3(1));
}
return ob->runtime.bb;
}
bool BKE_pointcloud_attribute_required(const PointCloud * /*pointcloud*/, const char *name)
{
return STREQ(name, POINTCLOUD_ATTR_POSITION);
}
/* Dependency Graph */
PointCloud *BKE_pointcloud_copy_for_eval(const PointCloud *pointcloud_src)
{
return reinterpret_cast<PointCloud *>(
BKE_id_copy_ex(nullptr, &pointcloud_src->id, nullptr, LIB_ID_COPY_LOCALIZE));
}
static void pointcloud_evaluate_modifiers(Depsgraph *depsgraph,
Scene *scene,
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
Object *object,
blender::bke::GeometrySet &geometry_set)
{
/* Modifier evaluation modes. */
const bool use_render = (DEG_get_mode(depsgraph) == DAG_EVAL_RENDER);
const int required_mode = use_render ? eModifierMode_Render : eModifierMode_Realtime;
ModifierApplyFlag apply_flag = use_render ? MOD_APPLY_RENDER : MOD_APPLY_USECACHE;
const ModifierEvalContext mectx = {depsgraph, object, apply_flag};
BKE_modifiers_clear_errors(object);
/* Get effective list of modifiers to execute. Some effects like shape keys
* are added as virtual modifiers before the user created modifiers. */
VirtualModifierData virtual_modifier_data;
ModifierData *md = BKE_modifiers_get_virtual_modifierlist(object, &virtual_modifier_data);
/* Evaluate modifiers. */
for (; md; md = md->next) {
const ModifierTypeInfo *mti = BKE_modifier_get_info((ModifierType)md->type);
if (!BKE_modifier_is_enabled(scene, md, required_mode)) {
continue;
}
Modifiers: measure execution time and provide Python access The goal is to give technical artists the ability to optimize modifier usage and/or geometry node groups for performance. In the long term, it would be useful if Blender could provide its own UI to display profiling information to users. However, right now, there are too many open design questions making it infeasible to tackle this in the short term. This commit uses a simpler approach: Instead of adding new ui for profiling data, it exposes the execution-time of modifiers in the Python API. This allows technical artists to access the information and to build their own UI to display the relevant information. In the long term this will hopefully also help us to integrate a native ui for this in Blender by observing how users use this information. Note: The execution time of a modifier highly depends on what other things the CPU is doing at the same time. For example, in many more complex files, many objects and therefore modifiers are evaluated at the same time by multiple threads which makes the measurement much less reliable. For best results, make sure that only one object is evaluated at a time (e.g. by changing it in isolation) and that no other process on the system keeps the CPU busy. As shown below, the execution time has to be accessed on the evaluated object, not the original object. ```lang=python import bpy depsgraph = bpy.context.view_layer.depsgraph ob = bpy.context.active_object ob_eval = ob.evaluated_get(depsgraph) modifier_eval = ob_eval.modifiers[0] print(modifier_eval.execution_time, "s") ``` Differential Revision: https://developer.blender.org/D17185
2023-02-06 15:39:59 +01:00
blender::bke::ScopedModifierTimer modifier_timer{*md};
if (mti->modify_geometry_set) {
mti->modify_geometry_set(md, &mectx, &geometry_set);
}
}
}
static PointCloud *take_pointcloud_ownership_from_geometry_set(
blender::bke::GeometrySet &geometry_set)
Geometry Nodes: support evaluating mesh object to geometry set This implements the design proposed in T83357. The goal is to allow the geometry nodes modifier on mesh objects to output instances and potentially other geometry types. Both problems are tackled by allowing mesh objects to evaluate to a geometry set, instead of just a single mesh id data block. The geometry set can contain a mesh but also other data like instances and a point cloud. I can't say that I'm sure that this commit won't introduce bugs. Mainly the temporary object creation during rendering seems a bit brittle. BUT, we can be reasonably sure that this commit will not introduce regressions (at least not ones, that are hard to fix). This is because the code has been written in a way that minimizes changes for existing functionality. Given that we intend to hide the point cloud object for the next release, we won't even have to worry about temporary object creation for now. An important part of the technical design is to make sure that `ObjectRuntime->data_eval` contains the same data before and after this patch. This helps to make sure, that existing code paths are impacted as little as possible. Instead of fully replacing `data_eval`, there is `geometry_set_eval`, which contains all the geometry components an object evaluated to (including the data referenced by `data_eval`). For now, not much code has to be aware of `geometry_set_eval`. Mainly the depsgraph object iterator and the instances system have to know about it. Reviewers: brecht Differential Revision: https://developer.blender.org/D9851
2020-12-15 12:42:10 +01:00
{
if (!geometry_set.has<blender::bke::PointCloudComponent>()) {
Geometry Nodes: support evaluating mesh object to geometry set This implements the design proposed in T83357. The goal is to allow the geometry nodes modifier on mesh objects to output instances and potentially other geometry types. Both problems are tackled by allowing mesh objects to evaluate to a geometry set, instead of just a single mesh id data block. The geometry set can contain a mesh but also other data like instances and a point cloud. I can't say that I'm sure that this commit won't introduce bugs. Mainly the temporary object creation during rendering seems a bit brittle. BUT, we can be reasonably sure that this commit will not introduce regressions (at least not ones, that are hard to fix). This is because the code has been written in a way that minimizes changes for existing functionality. Given that we intend to hide the point cloud object for the next release, we won't even have to worry about temporary object creation for now. An important part of the technical design is to make sure that `ObjectRuntime->data_eval` contains the same data before and after this patch. This helps to make sure, that existing code paths are impacted as little as possible. Instead of fully replacing `data_eval`, there is `geometry_set_eval`, which contains all the geometry components an object evaluated to (including the data referenced by `data_eval`). For now, not much code has to be aware of `geometry_set_eval`. Mainly the depsgraph object iterator and the instances system have to know about it. Reviewers: brecht Differential Revision: https://developer.blender.org/D9851
2020-12-15 12:42:10 +01:00
return nullptr;
}
blender::bke::PointCloudComponent &pointcloud_component =
geometry_set.get_component_for_write<blender::bke::PointCloudComponent>();
Geometry Nodes: support evaluating mesh object to geometry set This implements the design proposed in T83357. The goal is to allow the geometry nodes modifier on mesh objects to output instances and potentially other geometry types. Both problems are tackled by allowing mesh objects to evaluate to a geometry set, instead of just a single mesh id data block. The geometry set can contain a mesh but also other data like instances and a point cloud. I can't say that I'm sure that this commit won't introduce bugs. Mainly the temporary object creation during rendering seems a bit brittle. BUT, we can be reasonably sure that this commit will not introduce regressions (at least not ones, that are hard to fix). This is because the code has been written in a way that minimizes changes for existing functionality. Given that we intend to hide the point cloud object for the next release, we won't even have to worry about temporary object creation for now. An important part of the technical design is to make sure that `ObjectRuntime->data_eval` contains the same data before and after this patch. This helps to make sure, that existing code paths are impacted as little as possible. Instead of fully replacing `data_eval`, there is `geometry_set_eval`, which contains all the geometry components an object evaluated to (including the data referenced by `data_eval`). For now, not much code has to be aware of `geometry_set_eval`. Mainly the depsgraph object iterator and the instances system have to know about it. Reviewers: brecht Differential Revision: https://developer.blender.org/D9851
2020-12-15 12:42:10 +01:00
PointCloud *pointcloud = pointcloud_component.release();
if (pointcloud != nullptr) {
/* Add back, but as read-only non-owning component. */
pointcloud_component.replace(pointcloud, blender::bke::GeometryOwnershipType::ReadOnly);
Geometry Nodes: support evaluating mesh object to geometry set This implements the design proposed in T83357. The goal is to allow the geometry nodes modifier on mesh objects to output instances and potentially other geometry types. Both problems are tackled by allowing mesh objects to evaluate to a geometry set, instead of just a single mesh id data block. The geometry set can contain a mesh but also other data like instances and a point cloud. I can't say that I'm sure that this commit won't introduce bugs. Mainly the temporary object creation during rendering seems a bit brittle. BUT, we can be reasonably sure that this commit will not introduce regressions (at least not ones, that are hard to fix). This is because the code has been written in a way that minimizes changes for existing functionality. Given that we intend to hide the point cloud object for the next release, we won't even have to worry about temporary object creation for now. An important part of the technical design is to make sure that `ObjectRuntime->data_eval` contains the same data before and after this patch. This helps to make sure, that existing code paths are impacted as little as possible. Instead of fully replacing `data_eval`, there is `geometry_set_eval`, which contains all the geometry components an object evaluated to (including the data referenced by `data_eval`). For now, not much code has to be aware of `geometry_set_eval`. Mainly the depsgraph object iterator and the instances system have to know about it. Reviewers: brecht Differential Revision: https://developer.blender.org/D9851
2020-12-15 12:42:10 +01:00
}
else {
/* The component was empty, we can also remove it. */
geometry_set.remove<blender::bke::PointCloudComponent>();
Geometry Nodes: support evaluating mesh object to geometry set This implements the design proposed in T83357. The goal is to allow the geometry nodes modifier on mesh objects to output instances and potentially other geometry types. Both problems are tackled by allowing mesh objects to evaluate to a geometry set, instead of just a single mesh id data block. The geometry set can contain a mesh but also other data like instances and a point cloud. I can't say that I'm sure that this commit won't introduce bugs. Mainly the temporary object creation during rendering seems a bit brittle. BUT, we can be reasonably sure that this commit will not introduce regressions (at least not ones, that are hard to fix). This is because the code has been written in a way that minimizes changes for existing functionality. Given that we intend to hide the point cloud object for the next release, we won't even have to worry about temporary object creation for now. An important part of the technical design is to make sure that `ObjectRuntime->data_eval` contains the same data before and after this patch. This helps to make sure, that existing code paths are impacted as little as possible. Instead of fully replacing `data_eval`, there is `geometry_set_eval`, which contains all the geometry components an object evaluated to (including the data referenced by `data_eval`). For now, not much code has to be aware of `geometry_set_eval`. Mainly the depsgraph object iterator and the instances system have to know about it. Reviewers: brecht Differential Revision: https://developer.blender.org/D9851
2020-12-15 12:42:10 +01:00
}
return pointcloud;
}
void BKE_pointcloud_data_update(Depsgraph *depsgraph, Scene *scene, Object *object)
{
/* Free any evaluated data and restore original data. */
BKE_object_free_derived_caches(object);
/* Evaluate modifiers. */
PointCloud *pointcloud = static_cast<PointCloud *>(object->data);
blender::bke::GeometrySet geometry_set = blender::bke::GeometrySet::from_pointcloud(
pointcloud, blender::bke::GeometryOwnershipType::ReadOnly);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
pointcloud_evaluate_modifiers(depsgraph, scene, object, geometry_set);
Geometry Nodes: support evaluating mesh object to geometry set This implements the design proposed in T83357. The goal is to allow the geometry nodes modifier on mesh objects to output instances and potentially other geometry types. Both problems are tackled by allowing mesh objects to evaluate to a geometry set, instead of just a single mesh id data block. The geometry set can contain a mesh but also other data like instances and a point cloud. I can't say that I'm sure that this commit won't introduce bugs. Mainly the temporary object creation during rendering seems a bit brittle. BUT, we can be reasonably sure that this commit will not introduce regressions (at least not ones, that are hard to fix). This is because the code has been written in a way that minimizes changes for existing functionality. Given that we intend to hide the point cloud object for the next release, we won't even have to worry about temporary object creation for now. An important part of the technical design is to make sure that `ObjectRuntime->data_eval` contains the same data before and after this patch. This helps to make sure, that existing code paths are impacted as little as possible. Instead of fully replacing `data_eval`, there is `geometry_set_eval`, which contains all the geometry components an object evaluated to (including the data referenced by `data_eval`). For now, not much code has to be aware of `geometry_set_eval`. Mainly the depsgraph object iterator and the instances system have to know about it. Reviewers: brecht Differential Revision: https://developer.blender.org/D9851
2020-12-15 12:42:10 +01:00
PointCloud *pointcloud_eval = take_pointcloud_ownership_from_geometry_set(geometry_set);
/* If the geometry set did not contain a point cloud, we still create an empty one. */
if (pointcloud_eval == nullptr) {
pointcloud_eval = BKE_pointcloud_new_nomain(0);
}
/* Assign evaluated object. */
Geometry Nodes: support evaluating mesh object to geometry set This implements the design proposed in T83357. The goal is to allow the geometry nodes modifier on mesh objects to output instances and potentially other geometry types. Both problems are tackled by allowing mesh objects to evaluate to a geometry set, instead of just a single mesh id data block. The geometry set can contain a mesh but also other data like instances and a point cloud. I can't say that I'm sure that this commit won't introduce bugs. Mainly the temporary object creation during rendering seems a bit brittle. BUT, we can be reasonably sure that this commit will not introduce regressions (at least not ones, that are hard to fix). This is because the code has been written in a way that minimizes changes for existing functionality. Given that we intend to hide the point cloud object for the next release, we won't even have to worry about temporary object creation for now. An important part of the technical design is to make sure that `ObjectRuntime->data_eval` contains the same data before and after this patch. This helps to make sure, that existing code paths are impacted as little as possible. Instead of fully replacing `data_eval`, there is `geometry_set_eval`, which contains all the geometry components an object evaluated to (including the data referenced by `data_eval`). For now, not much code has to be aware of `geometry_set_eval`. Mainly the depsgraph object iterator and the instances system have to know about it. Reviewers: brecht Differential Revision: https://developer.blender.org/D9851
2020-12-15 12:42:10 +01:00
const bool eval_is_owned = pointcloud_eval != pointcloud;
BKE_object_eval_assign_data(object, &pointcloud_eval->id, eval_is_owned);
object->runtime.geometry_set_eval = new blender::bke::GeometrySet(std::move(geometry_set));
}
Geometry: Cache bounds min and max, share between data-blocks Bounding box calculation can be a large in some situations, especially instancing. This patch caches the min and max of the bounding box in runtime data of meshes, point clouds, and curves, implementing part of T96968. Bounds are now calculated lazily-- only after they are tagged dirty. Also, cached bounds are also shared when copying geometry data-blocks that have equivalent data. When bounds are calculated on an evaluated data-block, they are also accessible on the original, and the next evaluated ID will also share them. A geometry will stop sharing bounds as soon as its positions (or radii) are changed. Just caching the bounds gave a 2-3x speedup with thousands of mesh geometry instances in the viewport. Sharing the bounds can eliminate recalculations entirely in cases like copying meshes in geometry nodes or the selection paint brush in curves sculpt mode, which causes a reevaluation but doesn't change the positions. **Implementation** The sharing is achieved with a `shared_ptr` that points to a cache mutex (from D16419) and the cached bounds data. When geometries are copied, the bounds are shared by default, and only "un-shared" when the bounds are tagged dirty. Point clouds have a new runtime struct to store this data. Functions for tagging the data dirty are improved for added for point clouds and improved for curves. A missing tag has also been fixed for mesh sculpt mode. **Future** There are further improvements which can be worked on next - Apply changes to volume objects and other types where it makes sense - Continue cleanup changes described in T96968 - Apply shared cache design to more expensive data like triangulation or normals Differential Revision: https://developer.blender.org/D16204
2022-11-15 13:46:55 -06:00
void PointCloud::tag_positions_changed()
{
this->runtime->bounds_cache.tag_dirty();
}
void PointCloud::tag_radii_changed()
{
this->runtime->bounds_cache.tag_dirty();
}
/* Draw Cache */
void (*BKE_pointcloud_batch_cache_dirty_tag_cb)(PointCloud *pointcloud, int mode) = nullptr;
void (*BKE_pointcloud_batch_cache_free_cb)(PointCloud *pointcloud) = nullptr;
void BKE_pointcloud_batch_cache_dirty_tag(PointCloud *pointcloud, int mode)
{
if (pointcloud->batch_cache) {
BKE_pointcloud_batch_cache_dirty_tag_cb(pointcloud, mode);
}
}
void BKE_pointcloud_batch_cache_free(PointCloud *pointcloud)
{
if (pointcloud->batch_cache) {
BKE_pointcloud_batch_cache_free_cb(pointcloud);
}
}