**Changes**
As described in T93602, this patch removes all use of the `MVert`
struct, replacing it with a generic named attribute with the name
`"position"`, consistent with other geometry types.
Variable names have been changed from `verts` to `positions`, to align
with the attribute name and the more generic design (positions are not
vertices, they are just an attribute stored on the point domain).
This change is made possible by previous commits that moved all other
data out of `MVert` to runtime data or other generic attributes. What
remains is mostly a simple type change. Though, the type still shows up
859 times, so the patch is quite large.
One compromise is that now `CD_MASK_BAREMESH` now contains
`CD_PROP_FLOAT3`. With the general move towards generic attributes
over custom data types, we are removing use of these type masks anyway.
**Benefits**
The most obvious benefit is reduced memory usage and the benefits
that brings in memory-bound situations. `float3` is only 3 bytes, in
comparison to `MVert` which was 4. When there are millions of vertices
this starts to matter more.
The other benefits come from using a more generic type. Instead of
writing algorithms specifically for `MVert`, code can just use arrays
of vectors. This will allow eliminating many temporary arrays or
wrappers used to extract positions.
Many possible improvements aren't implemented in this patch, though
I did switch simplify or remove the process of creating temporary
position arrays in a few places.
The design clarity that "positions are just another attribute" brings
allows removing explicit copying of vertices in some procedural
operations-- they are just processed like most other attributes.
**Performance**
This touches so many areas that it's hard to benchmark exhaustively,
but I observed some areas as examples.
* The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster.
* The Spring splash screen went from ~4.3 to ~4.5 fps.
* The subdivision surface modifier/node was slightly faster
RNA access through Python may be slightly slower, since now we need
a name lookup instead of just a custom data type lookup for each index.
**Future Improvements**
* Remove uses of "vert_coords" functions:
* `BKE_mesh_vert_coords_alloc`
* `BKE_mesh_vert_coords_get`
* `BKE_mesh_vert_coords_apply{_with_mat4}`
* Remove more hidden copying of positions
* General simplification now possible in many areas
* Convert more code to C++ to use `float3` instead of `float[3]`
* Currently `reinterpret_cast` is used for those C-API functions
Differential Revision: https://developer.blender.org/D15982
357 lines
9.6 KiB
C
357 lines
9.6 KiB
C
/* SPDX-License-Identifier: GPL-2.0-or-later */
|
|
|
|
/** \file
|
|
* \ingroup edmesh
|
|
*
|
|
* Mirror calculation for edit-mode and object mode.
|
|
*/
|
|
|
|
#include "MEM_guardedalloc.h"
|
|
|
|
#include "BLI_math.h"
|
|
|
|
#include "DNA_mesh_types.h"
|
|
#include "DNA_meshdata_types.h"
|
|
#include "DNA_object_types.h"
|
|
|
|
#include "BKE_editmesh.h"
|
|
#include "BKE_mesh.h"
|
|
#include "BLI_kdtree.h"
|
|
|
|
#include "ED_mesh.h"
|
|
|
|
/* -------------------------------------------------------------------- */
|
|
/** \name Mesh Spatial Mirror API
|
|
* \{ */
|
|
|
|
#define KD_THRESH 0.00002f
|
|
|
|
static struct {
|
|
void *tree;
|
|
} MirrKdStore = {NULL};
|
|
|
|
void ED_mesh_mirror_spatial_table_begin(Object *ob, BMEditMesh *em, Mesh *me_eval)
|
|
{
|
|
Mesh *me = ob->data;
|
|
const bool use_em = (!me_eval && em && me->edit_mesh == em);
|
|
const int totvert = use_em ? em->bm->totvert : me_eval ? me_eval->totvert : me->totvert;
|
|
|
|
if (MirrKdStore.tree) { /* happens when entering this call without ending it */
|
|
ED_mesh_mirror_spatial_table_end(ob);
|
|
}
|
|
|
|
MirrKdStore.tree = BLI_kdtree_3d_new(totvert);
|
|
|
|
if (use_em) {
|
|
BMVert *eve;
|
|
BMIter iter;
|
|
int i;
|
|
|
|
/* this needs to be valid for index lookups later (callers need) */
|
|
BM_mesh_elem_table_ensure(em->bm, BM_VERT);
|
|
|
|
BM_ITER_MESH_INDEX (eve, &iter, em->bm, BM_VERTS_OF_MESH, i) {
|
|
BLI_kdtree_3d_insert(MirrKdStore.tree, i, eve->co);
|
|
}
|
|
}
|
|
else {
|
|
const float(*positions)[3] = BKE_mesh_vert_positions(me_eval ? me_eval : me);
|
|
for (int i = 0; i < totvert; i++) {
|
|
BLI_kdtree_3d_insert(MirrKdStore.tree, i, positions[i]);
|
|
}
|
|
}
|
|
|
|
BLI_kdtree_3d_balance(MirrKdStore.tree);
|
|
}
|
|
|
|
int ED_mesh_mirror_spatial_table_lookup(Object *ob,
|
|
BMEditMesh *em,
|
|
Mesh *me_eval,
|
|
const float co[3])
|
|
{
|
|
if (MirrKdStore.tree == NULL) {
|
|
ED_mesh_mirror_spatial_table_begin(ob, em, me_eval);
|
|
}
|
|
|
|
if (MirrKdStore.tree) {
|
|
KDTreeNearest_3d nearest;
|
|
const int i = BLI_kdtree_3d_find_nearest(MirrKdStore.tree, co, &nearest);
|
|
|
|
if (i != -1) {
|
|
if (nearest.dist < KD_THRESH) {
|
|
return i;
|
|
}
|
|
}
|
|
}
|
|
return -1;
|
|
}
|
|
|
|
void ED_mesh_mirror_spatial_table_end(Object *UNUSED(ob))
|
|
{
|
|
/* TODO: store this in object/object-data (keep unused argument for now). */
|
|
if (MirrKdStore.tree) {
|
|
BLI_kdtree_3d_free(MirrKdStore.tree);
|
|
MirrKdStore.tree = NULL;
|
|
}
|
|
}
|
|
|
|
/** \} */
|
|
|
|
/* -------------------------------------------------------------------- */
|
|
/** \name Mesh Topology Mirror API
|
|
* \{ */
|
|
|
|
typedef uint MirrTopoHash_t;
|
|
|
|
typedef struct MirrTopoVert_t {
|
|
MirrTopoHash_t hash;
|
|
int v_index;
|
|
} MirrTopoVert_t;
|
|
|
|
static int mirrtopo_hash_sort(const void *l1, const void *l2)
|
|
{
|
|
if ((MirrTopoHash_t)(intptr_t)l1 > (MirrTopoHash_t)(intptr_t)l2) {
|
|
return 1;
|
|
}
|
|
if ((MirrTopoHash_t)(intptr_t)l1 < (MirrTopoHash_t)(intptr_t)l2) {
|
|
return -1;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
static int mirrtopo_vert_sort(const void *v1, const void *v2)
|
|
{
|
|
if (((MirrTopoVert_t *)v1)->hash > ((MirrTopoVert_t *)v2)->hash) {
|
|
return 1;
|
|
}
|
|
if (((MirrTopoVert_t *)v1)->hash < ((MirrTopoVert_t *)v2)->hash) {
|
|
return -1;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
bool ED_mesh_mirrtopo_recalc_check(BMEditMesh *em, Mesh *me, MirrTopoStore_t *mesh_topo_store)
|
|
{
|
|
const bool is_editmode = em != NULL;
|
|
int totvert;
|
|
int totedge;
|
|
|
|
if (em) {
|
|
totvert = em->bm->totvert;
|
|
totedge = em->bm->totedge;
|
|
}
|
|
else {
|
|
totvert = me->totvert;
|
|
totedge = me->totedge;
|
|
}
|
|
|
|
if ((mesh_topo_store->index_lookup == NULL) ||
|
|
(mesh_topo_store->prev_is_editmode != is_editmode) ||
|
|
(totvert != mesh_topo_store->prev_vert_tot) || (totedge != mesh_topo_store->prev_edge_tot)) {
|
|
return true;
|
|
}
|
|
return false;
|
|
}
|
|
|
|
void ED_mesh_mirrtopo_init(BMEditMesh *em,
|
|
Mesh *me,
|
|
MirrTopoStore_t *mesh_topo_store,
|
|
const bool skip_em_vert_array_init)
|
|
{
|
|
if (em) {
|
|
BLI_assert(me == NULL);
|
|
}
|
|
const bool is_editmode = (em != NULL);
|
|
const MEdge *medge = NULL, *med;
|
|
|
|
/* Edit-mode variables. */
|
|
BMEdge *eed;
|
|
BMIter iter;
|
|
|
|
int a, last;
|
|
int totvert, totedge;
|
|
int tot_unique = -1, tot_unique_prev = -1;
|
|
int tot_unique_edges = 0, tot_unique_edges_prev;
|
|
|
|
MirrTopoHash_t *topo_hash = NULL;
|
|
MirrTopoHash_t *topo_hash_prev = NULL;
|
|
MirrTopoVert_t *topo_pairs;
|
|
MirrTopoHash_t topo_pass = 1;
|
|
|
|
intptr_t *index_lookup; /* direct access to mesh_topo_store->index_lookup */
|
|
|
|
/* reallocate if needed */
|
|
ED_mesh_mirrtopo_free(mesh_topo_store);
|
|
|
|
mesh_topo_store->prev_is_editmode = is_editmode;
|
|
|
|
if (em) {
|
|
BM_mesh_elem_index_ensure(em->bm, BM_VERT);
|
|
|
|
totvert = em->bm->totvert;
|
|
}
|
|
else {
|
|
totvert = me->totvert;
|
|
}
|
|
|
|
topo_hash = MEM_callocN(totvert * sizeof(MirrTopoHash_t), "TopoMirr");
|
|
|
|
/* Initialize the vert-edge-user counts used to detect unique topology */
|
|
if (em) {
|
|
totedge = em->bm->totedge;
|
|
|
|
BM_ITER_MESH (eed, &iter, em->bm, BM_EDGES_OF_MESH) {
|
|
const int i1 = BM_elem_index_get(eed->v1), i2 = BM_elem_index_get(eed->v2);
|
|
topo_hash[i1]++;
|
|
topo_hash[i2]++;
|
|
}
|
|
}
|
|
else {
|
|
totedge = me->totedge;
|
|
medge = BKE_mesh_edges(me);
|
|
for (a = 0, med = medge; a < totedge; a++, med++) {
|
|
const uint i1 = med->v1, i2 = med->v2;
|
|
topo_hash[i1]++;
|
|
topo_hash[i2]++;
|
|
}
|
|
}
|
|
|
|
topo_hash_prev = MEM_dupallocN(topo_hash);
|
|
|
|
tot_unique_prev = -1;
|
|
tot_unique_edges_prev = -1;
|
|
while (1) {
|
|
/* use the number of edges per vert to give verts unique topology IDs */
|
|
|
|
tot_unique_edges = 0;
|
|
|
|
/* This can make really big numbers, wrapping around here is fine */
|
|
if (em) {
|
|
BM_ITER_MESH (eed, &iter, em->bm, BM_EDGES_OF_MESH) {
|
|
const int i1 = BM_elem_index_get(eed->v1), i2 = BM_elem_index_get(eed->v2);
|
|
topo_hash[i1] += topo_hash_prev[i2] * topo_pass;
|
|
topo_hash[i2] += topo_hash_prev[i1] * topo_pass;
|
|
tot_unique_edges += (topo_hash[i1] != topo_hash[i2]);
|
|
}
|
|
}
|
|
else {
|
|
for (a = 0, med = medge; a < totedge; a++, med++) {
|
|
const uint i1 = med->v1, i2 = med->v2;
|
|
topo_hash[i1] += topo_hash_prev[i2] * topo_pass;
|
|
topo_hash[i2] += topo_hash_prev[i1] * topo_pass;
|
|
tot_unique_edges += (topo_hash[i1] != topo_hash[i2]);
|
|
}
|
|
}
|
|
memcpy(topo_hash_prev, topo_hash, sizeof(MirrTopoHash_t) * totvert);
|
|
|
|
/* sort so we can count unique values */
|
|
qsort(topo_hash_prev, totvert, sizeof(MirrTopoHash_t), mirrtopo_hash_sort);
|
|
|
|
tot_unique = 1; /* account for skipping the first value */
|
|
for (a = 1; a < totvert; a++) {
|
|
if (topo_hash_prev[a - 1] != topo_hash_prev[a]) {
|
|
tot_unique++;
|
|
}
|
|
}
|
|
|
|
if ((tot_unique <= tot_unique_prev) && (tot_unique_edges <= tot_unique_edges_prev)) {
|
|
/* Finish searching for unique values when 1 loop doesn't give a
|
|
* higher number of unique values compared to the previous loop. */
|
|
break;
|
|
}
|
|
tot_unique_prev = tot_unique;
|
|
tot_unique_edges_prev = tot_unique_edges;
|
|
/* Copy the hash calculated this iteration, so we can use them next time */
|
|
memcpy(topo_hash_prev, topo_hash, sizeof(MirrTopoHash_t) * totvert);
|
|
|
|
topo_pass++;
|
|
}
|
|
|
|
/* Hash/Index pairs are needed for sorting to find index pairs */
|
|
topo_pairs = MEM_callocN(sizeof(MirrTopoVert_t) * totvert, "MirrTopoPairs");
|
|
|
|
/* since we are looping through verts, initialize these values here too */
|
|
index_lookup = MEM_mallocN(totvert * sizeof(*index_lookup), "mesh_topo_lookup");
|
|
|
|
if (em) {
|
|
if (skip_em_vert_array_init == false) {
|
|
BM_mesh_elem_table_ensure(em->bm, BM_VERT);
|
|
}
|
|
}
|
|
|
|
for (a = 0; a < totvert; a++) {
|
|
topo_pairs[a].hash = topo_hash[a];
|
|
topo_pairs[a].v_index = a;
|
|
|
|
/* initialize lookup */
|
|
index_lookup[a] = -1;
|
|
}
|
|
|
|
qsort(topo_pairs, totvert, sizeof(MirrTopoVert_t), mirrtopo_vert_sort);
|
|
|
|
last = 0;
|
|
|
|
/* Get the pairs out of the sorted hashes.
|
|
* NOTE: `totvert + 1` means we can use the previous 2,
|
|
* but you can't ever access the last 'a' index of #MirrTopoPairs. */
|
|
if (em) {
|
|
BMVert **vtable = em->bm->vtable;
|
|
for (a = 1; a <= totvert; a++) {
|
|
// printf("I %d %ld %d\n",
|
|
// (a - last), MirrTopoPairs[a].hash, MirrTopoPairs[a].v_index);
|
|
if ((a == totvert) || (topo_pairs[a - 1].hash != topo_pairs[a].hash)) {
|
|
const int match_count = a - last;
|
|
if (match_count == 2) {
|
|
const int j = topo_pairs[a - 1].v_index, k = topo_pairs[a - 2].v_index;
|
|
index_lookup[j] = (intptr_t)vtable[k];
|
|
index_lookup[k] = (intptr_t)vtable[j];
|
|
}
|
|
else if (match_count == 1) {
|
|
/* Center vertex. */
|
|
const int j = topo_pairs[a - 1].v_index;
|
|
index_lookup[j] = (intptr_t)vtable[j];
|
|
}
|
|
last = a;
|
|
}
|
|
}
|
|
}
|
|
else {
|
|
/* same as above, for mesh */
|
|
for (a = 1; a <= totvert; a++) {
|
|
if ((a == totvert) || (topo_pairs[a - 1].hash != topo_pairs[a].hash)) {
|
|
const int match_count = a - last;
|
|
if (match_count == 2) {
|
|
const int j = topo_pairs[a - 1].v_index, k = topo_pairs[a - 2].v_index;
|
|
index_lookup[j] = k;
|
|
index_lookup[k] = j;
|
|
}
|
|
else if (match_count == 1) {
|
|
/* Center vertex. */
|
|
const int j = topo_pairs[a - 1].v_index;
|
|
index_lookup[j] = j;
|
|
}
|
|
last = a;
|
|
}
|
|
}
|
|
}
|
|
|
|
MEM_freeN(topo_pairs);
|
|
topo_pairs = NULL;
|
|
|
|
MEM_freeN(topo_hash);
|
|
MEM_freeN(topo_hash_prev);
|
|
|
|
mesh_topo_store->index_lookup = index_lookup;
|
|
mesh_topo_store->prev_vert_tot = totvert;
|
|
mesh_topo_store->prev_edge_tot = totedge;
|
|
}
|
|
|
|
void ED_mesh_mirrtopo_free(MirrTopoStore_t *mesh_topo_store)
|
|
{
|
|
MEM_SAFE_FREE(mesh_topo_store->index_lookup);
|
|
mesh_topo_store->prev_vert_tot = -1;
|
|
mesh_topo_store->prev_edge_tot = -1;
|
|
}
|
|
|
|
/** \} */
|