This repository has been archived on 2023-10-09. You can view files and clone it. You cannot open issues or pull requests or push a commit.
Files
blender-archive/source/blender/blenkernel/BKE_pbvh.h
Hans Goudey cfa53e0fbe Refactor: Move normals out of MVert, lazy calculation
As described in T91186, this commit moves mesh vertex normals into a
contiguous array of float vectors in a custom data layer, how face
normals are currently stored.

The main interface is documented in `BKE_mesh.h`. Vertex and face
normals are now calculated on-demand and cached, retrieved with an
"ensure" function. Since the logical state of a mesh is now "has
normals when necessary", they can be retrieved from a `const` mesh.

The goal is to use on-demand calculation for all derived data, but
leave room for eager calculation for performance purposes (modifier
evaluation is threaded, but viewport data generation is not).

**Benefits**
This moves us closer to a SoA approach rather than the current AoS
paradigm. Accessing a contiguous `float3` is much more efficient than
retrieving data from a larger struct. The memory requirements for
accessing only normals or vertex locations are smaller, and at the
cost of more memory usage for just normals, they now don't have to
be converted between float and short, which also simplifies code

In the future, the remaining items can be removed from `MVert`,
leaving only `float3`, which has similar benefits (see T93602).

Removing the combination of derived and original data makes it
conceptually simpler to only calculate normals when necessary.
This is especially important now that we have more opportunities
for temporary meshes in geometry nodes.

**Performance**
In addition to the theoretical future performance improvements by
making `MVert == float3`, I've done some basic performance testing
on this patch directly. The data is fairly rough, but it gives an idea
about where things stand generally.
 - Mesh line primitive 4m Verts: 1.16x faster (36 -> 31 ms),
   showing that accessing just `MVert` is now more efficient.
 - Spring Splash Screen: 1.03-1.06 -> 1.06-1.11 FPS, a very slight
   change that at least shows there is no regression.
 - Sprite Fright Snail Smoosh: 3.30-3.40 -> 3.42-3.50 FPS, a small
   but observable speedup.
 - Set Position Node with Scaled Normal: 1.36x faster (53 -> 39 ms),
   shows that using normals in geometry nodes is faster.
 - Normal Calculation 1.6m Vert Cube: 1.19x faster (25 -> 21 ms),
   shows that calculating normals is slightly faster now.
 - File Size of 1.6m Vert Cube: 1.03x smaller (214.7 -> 208.4 MB),
   Normals are not saved in files, which can help with large meshes.

As for memory usage, it may be slightly more in some cases, but
I didn't observe any difference in the production files I tested.

**Tests**
Some modifiers and cycles test results need to be updated with this
commit, for two reasons:
 - The subdivision surface modifier is not responsible for calculating
   normals anymore. In master, the modifier creates different normals
   than the result of the `Mesh` normal calculation, so this is a bug
   fix.
 - There are small differences in the results of some modifiers that
   use normals because they are not converted to and from `short`
   anymore.

**Future improvements**
 - Remove `ModifierTypeInfo::dependsOnNormals`. Code in each modifier
   already retrieves normals if they are needed anyway.
 - Copy normals as part of a better CoW system for attributes.
 - Make more areas use lazy instead of eager normal calculation.
 - Remove `BKE_mesh_normals_tag_dirty` in more places since that is
   now the default state of a new mesh.
 - Possibly apply a similar change to derived face corner normals.

Differential Revision: https://developer.blender.org/D12770
2022-01-13 14:38:25 -06:00

545 lines
18 KiB
C++

/*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*/
#pragma once
/** \file
* \ingroup bke
* \brief A BVH for high poly meshes.
*/
#include "BLI_bitmap.h"
#include "BLI_ghash.h"
/* For embedding CCGKey in iterator. */
#include "BKE_ccg.h"
#ifdef __cplusplus
extern "C" {
#endif
struct BMLog;
struct BMesh;
struct CCGElem;
struct CCGKey;
struct CustomData;
struct DMFlagMat;
struct GPU_PBVH_Buffers;
struct IsectRayPrecalc;
struct MLoop;
struct MLoopTri;
struct MPoly;
struct MVert;
struct Mesh;
struct PBVH;
struct PBVHNode;
struct SubdivCCG;
struct TaskParallelSettings;
typedef struct PBVH PBVH;
typedef struct PBVHNode PBVHNode;
typedef struct {
float (*co)[3];
} PBVHProxyNode;
typedef struct {
float (*color)[4];
} PBVHColorBufferNode;
typedef enum {
PBVH_Leaf = 1 << 0,
PBVH_UpdateNormals = 1 << 1,
PBVH_UpdateBB = 1 << 2,
PBVH_UpdateOriginalBB = 1 << 3,
PBVH_UpdateDrawBuffers = 1 << 4,
PBVH_UpdateRedraw = 1 << 5,
PBVH_UpdateMask = 1 << 6,
PBVH_UpdateVisibility = 1 << 8,
PBVH_RebuildDrawBuffers = 1 << 9,
PBVH_FullyHidden = 1 << 10,
PBVH_FullyMasked = 1 << 11,
PBVH_FullyUnmasked = 1 << 12,
PBVH_UpdateTopology = 1 << 13,
PBVH_UpdateColor = 1 << 14,
} PBVHNodeFlags;
typedef struct PBVHFrustumPlanes {
float (*planes)[4];
int num_planes;
} PBVHFrustumPlanes;
void BKE_pbvh_set_frustum_planes(PBVH *pbvh, PBVHFrustumPlanes *planes);
void BKE_pbvh_get_frustum_planes(PBVH *pbvh, PBVHFrustumPlanes *planes);
/* Callbacks */
/**
* Returns true if the search should continue from this node, false otherwise.
*/
typedef bool (*BKE_pbvh_SearchCallback)(PBVHNode *node, void *data);
typedef void (*BKE_pbvh_HitCallback)(PBVHNode *node, void *data);
typedef void (*BKE_pbvh_HitOccludedCallback)(PBVHNode *node, void *data, float *tmin);
typedef void (*BKE_pbvh_SearchNearestCallback)(PBVHNode *node, void *data, float *tmin);
/* Building */
PBVH *BKE_pbvh_new(void);
/**
* Do a full rebuild with on Mesh data structure.
*
* \note Unlike mpoly/mloop/verts, looptri is *totally owned* by PBVH
* (which means it may rewrite it if needed, see #BKE_pbvh_vert_coords_apply().
*/
void BKE_pbvh_build_mesh(PBVH *pbvh,
struct Mesh *mesh,
const struct MPoly *mpoly,
const struct MLoop *mloop,
struct MVert *verts,
int totvert,
struct CustomData *vdata,
struct CustomData *ldata,
struct CustomData *pdata,
const struct MLoopTri *looptri,
int looptri_num);
/**
* Do a full rebuild with on Grids data structure.
*/
void BKE_pbvh_build_grids(PBVH *pbvh,
struct CCGElem **grids,
int totgrid,
struct CCGKey *key,
void **gridfaces,
struct DMFlagMat *flagmats,
unsigned int **grid_hidden);
/**
* Build a PBVH from a BMesh.
*/
void BKE_pbvh_build_bmesh(PBVH *pbvh,
struct BMesh *bm,
bool smooth_shading,
struct BMLog *log,
int cd_vert_node_offset,
int cd_face_node_offset);
void BKE_pbvh_free(PBVH *pbvh);
/* Hierarchical Search in the BVH, two methods:
* - For each hit calling a callback.
* - Gather nodes in an array (easy to multi-thread). */
void BKE_pbvh_search_callback(PBVH *pbvh,
BKE_pbvh_SearchCallback scb,
void *search_data,
BKE_pbvh_HitCallback hcb,
void *hit_data);
void BKE_pbvh_search_gather(
PBVH *pbvh, BKE_pbvh_SearchCallback scb, void *search_data, PBVHNode ***array, int *tot);
/* Ray-cast
* the hit callback is called for all leaf nodes intersecting the ray;
* it's up to the callback to find the primitive within the leaves that is
* hit first */
void BKE_pbvh_raycast(PBVH *pbvh,
BKE_pbvh_HitOccludedCallback cb,
void *data,
const float ray_start[3],
const float ray_normal[3],
bool original);
bool BKE_pbvh_node_raycast(PBVH *pbvh,
PBVHNode *node,
float (*origco)[3],
bool use_origco,
const float ray_start[3],
const float ray_normal[3],
struct IsectRayPrecalc *isect_precalc,
float *depth,
int *active_vertex_index,
int *active_face_grid_index,
float *face_normal);
bool BKE_pbvh_bmesh_node_raycast_detail(PBVHNode *node,
const float ray_start[3],
struct IsectRayPrecalc *isect_precalc,
float *depth,
float *r_edge_length);
/**
* For orthographic cameras, project the far away ray segment points to the root node so
* we can have better precision.
*/
void BKE_pbvh_raycast_project_ray_root(
PBVH *pbvh, bool original, float ray_start[3], float ray_end[3], float ray_normal[3]);
void BKE_pbvh_find_nearest_to_ray(PBVH *pbvh,
BKE_pbvh_HitOccludedCallback cb,
void *data,
const float ray_start[3],
const float ray_normal[3],
bool original);
bool BKE_pbvh_node_find_nearest_to_ray(PBVH *pbvh,
PBVHNode *node,
float (*origco)[3],
bool use_origco,
const float ray_start[3],
const float ray_normal[3],
float *depth,
float *dist_sq);
/* Drawing */
void BKE_pbvh_draw_cb(PBVH *pbvh,
bool update_only_visible,
PBVHFrustumPlanes *update_frustum,
PBVHFrustumPlanes *draw_frustum,
void (*draw_fn)(void *user_data, struct GPU_PBVH_Buffers *buffers),
void *user_data);
void BKE_pbvh_draw_debug_cb(
PBVH *pbvh,
void (*draw_fn)(void *user_data, const float bmin[3], const float bmax[3], PBVHNodeFlags flag),
void *user_data);
/* PBVH Access */
typedef enum {
PBVH_FACES,
PBVH_GRIDS,
PBVH_BMESH,
} PBVHType;
PBVHType BKE_pbvh_type(const PBVH *pbvh);
bool BKE_pbvh_has_faces(const PBVH *pbvh);
/**
* Get the PBVH root's bounding box.
*/
void BKE_pbvh_bounding_box(const PBVH *pbvh, float min[3], float max[3]);
/**
* Multi-res hidden data, only valid for type == PBVH_GRIDS.
*/
unsigned int **BKE_pbvh_grid_hidden(const PBVH *pbvh);
/**
* Returns the number of visible quads in the nodes' grids.
*/
int BKE_pbvh_count_grid_quads(BLI_bitmap **grid_hidden,
const int *grid_indices,
int totgrid,
int gridsize);
void BKE_pbvh_sync_face_sets_to_grids(PBVH *pbvh);
/**
* Multi-res level, only valid for type == #PBVH_GRIDS.
*/
const struct CCGKey *BKE_pbvh_get_grid_key(const PBVH *pbvh);
struct CCGElem **BKE_pbvh_get_grids(const PBVH *pbvh);
BLI_bitmap **BKE_pbvh_get_grid_visibility(const PBVH *pbvh);
int BKE_pbvh_get_grid_num_vertices(const PBVH *pbvh);
int BKE_pbvh_get_grid_num_faces(const PBVH *pbvh);
/**
* Only valid for type == #PBVH_BMESH.
*/
struct BMesh *BKE_pbvh_get_bmesh(PBVH *pbvh);
void BKE_pbvh_bmesh_detail_size_set(PBVH *pbvh, float detail_size);
typedef enum {
PBVH_Subdivide = 1,
PBVH_Collapse = 2,
} PBVHTopologyUpdateMode;
/**
* Collapse short edges, subdivide long edges.
*/
bool BKE_pbvh_bmesh_update_topology(PBVH *pbvh,
PBVHTopologyUpdateMode mode,
const float center[3],
const float view_normal[3],
float radius,
bool use_frontface,
bool use_projected);
/* Node Access */
void BKE_pbvh_node_mark_update(PBVHNode *node);
void BKE_pbvh_node_mark_update_mask(PBVHNode *node);
void BKE_pbvh_node_mark_update_color(PBVHNode *node);
void BKE_pbvh_node_mark_update_visibility(PBVHNode *node);
void BKE_pbvh_node_mark_rebuild_draw(PBVHNode *node);
void BKE_pbvh_node_mark_redraw(PBVHNode *node);
void BKE_pbvh_node_mark_normals_update(PBVHNode *node);
void BKE_pbvh_node_mark_topology_update(PBVHNode *node);
void BKE_pbvh_node_fully_hidden_set(PBVHNode *node, int fully_hidden);
bool BKE_pbvh_node_fully_hidden_get(PBVHNode *node);
void BKE_pbvh_node_fully_masked_set(PBVHNode *node, int fully_masked);
bool BKE_pbvh_node_fully_masked_get(PBVHNode *node);
void BKE_pbvh_node_fully_unmasked_set(PBVHNode *node, int fully_masked);
bool BKE_pbvh_node_fully_unmasked_get(PBVHNode *node);
void BKE_pbvh_node_get_grids(PBVH *pbvh,
PBVHNode *node,
int **grid_indices,
int *totgrid,
int *maxgrid,
int *gridsize,
struct CCGElem ***r_griddata);
void BKE_pbvh_node_num_verts(PBVH *pbvh, PBVHNode *node, int *r_uniquevert, int *r_totvert);
void BKE_pbvh_node_get_verts(PBVH *pbvh,
PBVHNode *node,
const int **r_vert_indices,
struct MVert **r_verts);
void BKE_pbvh_node_get_BB(PBVHNode *node, float bb_min[3], float bb_max[3]);
void BKE_pbvh_node_get_original_BB(PBVHNode *node, float bb_min[3], float bb_max[3]);
float BKE_pbvh_node_get_tmin(PBVHNode *node);
/**
* Test if AABB is at least partially inside the #PBVHFrustumPlanes volume.
*/
bool BKE_pbvh_node_frustum_contain_AABB(PBVHNode *node, void *frustum);
/**
* Test if AABB is at least partially outside the #PBVHFrustumPlanes volume.
*/
bool BKE_pbvh_node_frustum_exclude_AABB(PBVHNode *node, void *frustum);
struct GSet *BKE_pbvh_bmesh_node_unique_verts(PBVHNode *node);
struct GSet *BKE_pbvh_bmesh_node_other_verts(PBVHNode *node);
struct GSet *BKE_pbvh_bmesh_node_faces(PBVHNode *node);
/**
* In order to perform operations on the original node coordinates
* (currently just ray-cast), store the node's triangles and vertices.
*
* Skips triangles that are hidden.
*/
void BKE_pbvh_bmesh_node_save_orig(struct BMesh *bm, PBVHNode *node);
void BKE_pbvh_bmesh_after_stroke(PBVH *pbvh);
/* Update Bounding Box/Redraw and clear flags. */
void BKE_pbvh_update_bounds(PBVH *pbvh, int flags);
void BKE_pbvh_update_vertex_data(PBVH *pbvh, int flags);
void BKE_pbvh_update_visibility(PBVH *pbvh);
void BKE_pbvh_update_normals(PBVH *pbvh, struct SubdivCCG *subdiv_ccg);
void BKE_pbvh_redraw_BB(PBVH *pbvh, float bb_min[3], float bb_max[3]);
void BKE_pbvh_get_grid_updates(PBVH *pbvh, bool clear, void ***r_gridfaces, int *r_totface);
void BKE_pbvh_grids_update(PBVH *pbvh,
struct CCGElem **grids,
void **gridfaces,
struct DMFlagMat *flagmats,
unsigned int **grid_hidden);
void BKE_pbvh_subdiv_cgg_set(PBVH *pbvh, struct SubdivCCG *subdiv_ccg);
void BKE_pbvh_face_sets_set(PBVH *pbvh, int *face_sets);
void BKE_pbvh_face_sets_color_set(PBVH *pbvh, int seed, int color_default);
void BKE_pbvh_respect_hide_set(PBVH *pbvh, bool respect_hide);
/* Vertex Deformer. */
float (*BKE_pbvh_vert_coords_alloc(struct PBVH *pbvh))[3];
void BKE_pbvh_vert_coords_apply(struct PBVH *pbvh, const float (*vertCos)[3], int totvert);
bool BKE_pbvh_is_deformed(struct PBVH *pbvh);
/* Vertex Iterator. */
/* This iterator has quite a lot of code, but it's designed to:
* - allow the compiler to eliminate dead code and variables
* - spend most of the time in the relatively simple inner loop */
/* NOTE: PBVH_ITER_ALL does not skip hidden vertices,
* PBVH_ITER_UNIQUE does */
#define PBVH_ITER_ALL 0
#define PBVH_ITER_UNIQUE 1
typedef struct PBVHVertexIter {
/* iteration */
int g;
int width;
int height;
int gx;
int gy;
int i;
int index;
bool respect_hide;
/* grid */
struct CCGKey key;
struct CCGElem **grids;
struct CCGElem *grid;
BLI_bitmap **grid_hidden, *gh;
int *grid_indices;
int totgrid;
int gridsize;
/* mesh */
struct MVert *mverts;
float (*vert_normals)[3];
int totvert;
const int *vert_indices;
struct MPropCol *vcol;
float *vmask;
/* bmesh */
struct GSetIterator bm_unique_verts;
struct GSetIterator bm_other_verts;
struct CustomData *bm_vdata;
int cd_vert_mask_offset;
/* result: these are all computed in the macro, but we assume
* that compiler optimization's will skip the ones we don't use */
struct MVert *mvert;
struct BMVert *bm_vert;
float *co;
float *no;
float *fno;
float *mask;
float *col;
bool visible;
} PBVHVertexIter;
void pbvh_vertex_iter_init(PBVH *pbvh, PBVHNode *node, PBVHVertexIter *vi, int mode);
#define BKE_pbvh_vertex_iter_begin(pbvh, node, vi, mode) \
pbvh_vertex_iter_init(pbvh, node, &vi, mode); \
\
for (vi.i = 0, vi.g = 0; vi.g < vi.totgrid; vi.g++) { \
if (vi.grids) { \
vi.width = vi.gridsize; \
vi.height = vi.gridsize; \
vi.index = vi.grid_indices[vi.g] * vi.key.grid_area - 1; \
vi.grid = vi.grids[vi.grid_indices[vi.g]]; \
if (mode == PBVH_ITER_UNIQUE) { \
vi.gh = vi.grid_hidden[vi.grid_indices[vi.g]]; \
} \
} \
else { \
vi.width = vi.totvert; \
vi.height = 1; \
} \
\
for (vi.gy = 0; vi.gy < vi.height; vi.gy++) { \
for (vi.gx = 0; vi.gx < vi.width; vi.gx++, vi.i++) { \
if (vi.grid) { \
vi.co = CCG_elem_co(&vi.key, vi.grid); \
vi.fno = CCG_elem_no(&vi.key, vi.grid); \
vi.mask = vi.key.has_mask ? CCG_elem_mask(&vi.key, vi.grid) : NULL; \
vi.grid = CCG_elem_next(&vi.key, vi.grid); \
vi.index++; \
vi.visible = true; \
if (vi.gh) { \
if (BLI_BITMAP_TEST(vi.gh, vi.gy * vi.gridsize + vi.gx)) { \
continue; \
} \
} \
} \
else if (vi.mverts) { \
vi.mvert = &vi.mverts[vi.vert_indices[vi.gx]]; \
if (vi.respect_hide) { \
vi.visible = !(vi.mvert->flag & ME_HIDE); \
if (mode == PBVH_ITER_UNIQUE && !vi.visible) { \
continue; \
} \
} \
else { \
BLI_assert(vi.visible); \
} \
vi.co = vi.mvert->co; \
vi.no = vi.vert_normals[vi.vert_indices[vi.gx]]; \
vi.index = vi.vert_indices[vi.i]; \
if (vi.vmask) { \
vi.mask = &vi.vmask[vi.index]; \
} \
if (vi.vcol) { \
vi.col = vi.vcol[vi.index].color; \
} \
} \
else { \
if (!BLI_gsetIterator_done(&vi.bm_unique_verts)) { \
vi.bm_vert = BLI_gsetIterator_getKey(&vi.bm_unique_verts); \
BLI_gsetIterator_step(&vi.bm_unique_verts); \
} \
else { \
vi.bm_vert = BLI_gsetIterator_getKey(&vi.bm_other_verts); \
BLI_gsetIterator_step(&vi.bm_other_verts); \
} \
vi.visible = !BM_elem_flag_test_bool(vi.bm_vert, BM_ELEM_HIDDEN); \
if (mode == PBVH_ITER_UNIQUE && !vi.visible) { \
continue; \
} \
vi.co = vi.bm_vert->co; \
vi.fno = vi.bm_vert->no; \
vi.index = BM_elem_index_get(vi.bm_vert); \
vi.mask = BM_ELEM_CD_GET_VOID_P(vi.bm_vert, vi.cd_vert_mask_offset); \
}
#define BKE_pbvh_vertex_iter_end \
} \
} \
} \
((void)0)
void BKE_pbvh_node_get_proxies(PBVHNode *node, PBVHProxyNode **proxies, int *proxy_count);
void BKE_pbvh_node_free_proxies(PBVHNode *node);
PBVHProxyNode *BKE_pbvh_node_add_proxy(PBVH *pbvh, PBVHNode *node);
void BKE_pbvh_gather_proxies(PBVH *pbvh, PBVHNode ***r_array, int *r_tot);
void BKE_pbvh_node_get_bm_orco_data(PBVHNode *node,
int (**r_orco_tris)[3],
int *r_orco_tris_num,
float (**r_orco_coords)[3]);
/**
* \note doing a full search on all vertices here seems expensive,
* however this is important to avoid having to recalculate bound-box & sync the buffers to the
* GPU (which is far more expensive!) See: T47232.
*/
bool BKE_pbvh_node_vert_update_check_any(PBVH *pbvh, PBVHNode *node);
// void BKE_pbvh_node_BB_reset(PBVHNode *node);
// void BKE_pbvh_node_BB_expand(PBVHNode *node, float co[3]);
bool pbvh_has_mask(const PBVH *pbvh);
void pbvh_show_mask_set(PBVH *pbvh, bool show_mask);
bool pbvh_has_face_sets(PBVH *pbvh);
void pbvh_show_face_sets_set(PBVH *pbvh, bool show_face_sets);
/* Parallelization. */
void BKE_pbvh_parallel_range_settings(struct TaskParallelSettings *settings,
bool use_threading,
int totnode);
struct MVert *BKE_pbvh_get_verts(const PBVH *pbvh);
const float (*BKE_pbvh_get_vert_normals(const PBVH *pbvh))[3];
PBVHColorBufferNode *BKE_pbvh_node_color_buffer_get(PBVHNode *node);
void BKE_pbvh_node_color_buffer_free(PBVH *pbvh);
#ifdef __cplusplus
}
#endif