Welcome to mirror list, hosted at ThFree Co, Russian Federation.

git.blender.org/blender.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorHans Goudey <h.goudey@me.com>2022-09-05 19:56:34 +0300
committerHans Goudey <h.goudey@me.com>2022-09-05 19:56:34 +0300
commit05952aa94d33eeb504fa63618ba35c2bcc8bd19b (patch)
treec9ec37adf20c3c37ccaab44869220dcbe8e987a3 /source/blender/blenkernel/intern/fluid.c
parent63cfc8f9f6d623f33b50c5c07976af2b22845713 (diff)
Mesh: Remove redundant custom data pointers
For copy-on-write, we want to share attribute arrays between meshes where possible. Mutable pointers like `Mesh.mvert` make that difficult by making ownership vague. They also make code more complex by adding redundancy. The simplest solution is just removing them and retrieving layers from `CustomData` as needed. Similar changes have already been applied to curves and point clouds (e9f82d3dc7ee, 410a6efb747f). Removing use of the pointers generally makes code more obvious and more reusable. Mesh data is now accessed with a C++ API (`Mesh::edges()` or `Mesh::edges_for_write()`), and a C API (`BKE_mesh_edges(mesh)`). The CoW changes this commit makes possible are described in T95845 and T95842, and started in D14139 and D14140. The change also simplifies the ongoing mesh struct-of-array refactors from T95965. **RNA/Python Access Performance** Theoretically, accessing mesh elements with the RNA API may become slower, since the layer needs to be found on every random access. However, overhead is already high enough that this doesn't make a noticible differenc, and performance is actually improved in some cases. Random access can be up to 10% faster, but other situations might be a bit slower. Generally using `foreach_get/set` are the best way to improve performance. See the differential revision for more discussion about Python performance. Cycles has been updated to use raw pointers and the internal Blender mesh types, mostly because there is no sense in having this overhead when it's already compiled with Blender. In my tests this roughly halves the Cycles mesh creation time (0.19s to 0.10s for a 1 million face grid). Differential Revision: https://developer.blender.org/D15488
Diffstat (limited to 'source/blender/blenkernel/intern/fluid.c')
-rw-r--r--source/blender/blenkernel/intern/fluid.c69
1 files changed, 28 insertions, 41 deletions
diff --git a/source/blender/blenkernel/intern/fluid.c b/source/blender/blenkernel/intern/fluid.c
index 74ff10cc32c..24c61a792eb 100644
--- a/source/blender/blenkernel/intern/fluid.c
+++ b/source/blender/blenkernel/intern/fluid.c
@@ -403,7 +403,8 @@ static void manta_set_domain_from_mesh(FluidDomainSettings *fds,
size_t i;
float min[3] = {FLT_MAX, FLT_MAX, FLT_MAX}, max[3] = {-FLT_MAX, -FLT_MAX, -FLT_MAX};
float size[3];
- MVert *verts = me->mvert;
+
+ MVert *verts = BKE_mesh_vertices_for_write(me);
float scale = 0.0;
int res;
@@ -994,9 +995,7 @@ static void obstacles_from_mesh(Object *coll_ob,
{
if (fes->mesh) {
Mesh *me = NULL;
- MVert *mvert = NULL;
const MLoopTri *looptri;
- const MLoop *mloop;
BVHTreeFromMesh tree_data = {NULL};
int numverts, i;
@@ -1008,13 +1007,9 @@ static void obstacles_from_mesh(Object *coll_ob,
int min[3], max[3], res[3];
/* Duplicate vertices to modify. */
- if (me->mvert) {
- me->mvert = MEM_dupallocN(me->mvert);
- CustomData_set_layer(&me->vdata, CD_MVERT, me->mvert);
- }
+ MVert *verts = MEM_dupallocN(BKE_mesh_vertices(me));
- mvert = me->mvert;
- mloop = me->mloop;
+ const MLoop *mloop = BKE_mesh_loops(me);
looptri = BKE_mesh_runtime_looptri_ensure(me);
numverts = me->totvert;
@@ -1041,11 +1036,11 @@ static void obstacles_from_mesh(Object *coll_ob,
float co[3];
/* Vertex position. */
- mul_m4_v3(coll_ob->obmat, mvert[i].co);
- manta_pos_to_cell(fds, mvert[i].co);
+ mul_m4_v3(coll_ob->obmat, verts[i].co);
+ manta_pos_to_cell(fds, verts[i].co);
/* Vertex velocity. */
- add_v3fl_v3fl_v3i(co, mvert[i].co, fds->shift);
+ add_v3fl_v3fl_v3i(co, verts[i].co, fds->shift);
if (has_velocity) {
sub_v3_v3v3(&vert_vel[i * 3], co, &fes->verts_old[i * 3]);
mul_v3_fl(&vert_vel[i * 3], 1.0f / dt);
@@ -1053,7 +1048,7 @@ static void obstacles_from_mesh(Object *coll_ob,
copy_v3_v3(&fes->verts_old[i * 3], co);
/* Calculate emission map bounds. */
- bb_boundInsert(bb, mvert[i].co);
+ bb_boundInsert(bb, verts[i].co);
}
/* Set emission map.
@@ -1075,7 +1070,7 @@ static void obstacles_from_mesh(Object *coll_ob,
ObstaclesFromDMData data = {
.fes = fes,
- .mvert = mvert,
+ .mvert = verts,
.mloop = mloop,
.mlooptri = looptri,
.tree = &tree_data,
@@ -1098,9 +1093,7 @@ static void obstacles_from_mesh(Object *coll_ob,
if (vert_vel) {
MEM_freeN(vert_vel);
}
- if (me->mvert) {
- MEM_freeN(me->mvert);
- }
+ MEM_SAFE_FREE(verts);
BKE_id_free(NULL, me);
}
}
@@ -2080,16 +2073,12 @@ static void emit_from_mesh(
Mesh *me = BKE_mesh_copy_for_eval(ffs->mesh, true);
/* Duplicate vertices to modify. */
- if (me->mvert) {
- me->mvert = MEM_dupallocN(me->mvert);
- CustomData_set_layer(&me->vdata, CD_MVERT, me->mvert);
- }
+ MVert *verts = MEM_dupallocN(BKE_mesh_vertices(me));
- MVert *mvert = me->mvert;
- const MLoop *mloop = me->mloop;
+ const MLoop *mloop = BKE_mesh_loops(me);
const MLoopTri *mlooptri = BKE_mesh_runtime_looptri_ensure(me);
const int numverts = me->totvert;
- const MDeformVert *dvert = CustomData_get_layer(&me->vdata, CD_MDEFORMVERT);
+ const MDeformVert *dvert = BKE_mesh_deform_verts(me);
const MLoopUV *mloopuv = CustomData_get_layer_named(&me->ldata, CD_MLOOPUV, ffs->uvlayer_name);
if (ffs->flags & FLUID_FLOW_INITVELOCITY) {
@@ -2109,12 +2098,11 @@ static void emit_from_mesh(
/* Transform mesh vertices to domain grid space for fast lookups.
* This is valid because the mesh is copied above. */
- BKE_mesh_vertex_normals_ensure(me);
- float(*vert_normals)[3] = BKE_mesh_vertex_normals_for_write(me);
+ float(*vert_normals)[3] = MEM_dupallocN(BKE_mesh_vertex_normals_ensure(me));
for (i = 0; i < numverts; i++) {
/* Vertex position. */
- mul_m4_v3(flow_ob->obmat, mvert[i].co);
- manta_pos_to_cell(fds, mvert[i].co);
+ mul_m4_v3(flow_ob->obmat, verts[i].co);
+ manta_pos_to_cell(fds, verts[i].co);
/* Vertex normal. */
mul_mat3_m4_v3(flow_ob->obmat, vert_normals[i]);
@@ -2124,7 +2112,7 @@ static void emit_from_mesh(
/* Vertex velocity. */
if (ffs->flags & FLUID_FLOW_INITVELOCITY) {
float co[3];
- add_v3fl_v3fl_v3i(co, mvert[i].co, fds->shift);
+ add_v3fl_v3fl_v3i(co, verts[i].co, fds->shift);
if (has_velocity) {
sub_v3_v3v3(&vert_vel[i * 3], co, &ffs->verts_old[i * 3]);
mul_v3_fl(&vert_vel[i * 3], 1.0 / dt);
@@ -2133,7 +2121,7 @@ static void emit_from_mesh(
}
/* Calculate emission map bounds. */
- bb_boundInsert(bb, mvert[i].co);
+ bb_boundInsert(bb, verts[i].co);
}
mul_m4_v3(flow_ob->obmat, flow_center);
manta_pos_to_cell(fds, flow_center);
@@ -2158,7 +2146,7 @@ static void emit_from_mesh(
EmitFromDMData data = {
.fds = fds,
.ffs = ffs,
- .mvert = mvert,
+ .mvert = verts,
.vert_normals = vert_normals,
.mloop = mloop,
.mlooptri = mlooptri,
@@ -2186,9 +2174,8 @@ static void emit_from_mesh(
if (vert_vel) {
MEM_freeN(vert_vel);
}
- if (me->mvert) {
- MEM_freeN(me->mvert);
- }
+ MEM_SAFE_FREE(verts);
+ MEM_SAFE_FREE(vert_normals);
BKE_id_free(NULL, me);
}
}
@@ -3241,7 +3228,7 @@ static Mesh *create_liquid_geometry(FluidDomainSettings *fds,
* If there are no faces in original mesh, keep materials and flags unchanged. */
MPoly *mpoly;
MPoly mp_example = {0};
- mpoly = orgmesh->mpoly;
+ mpoly = BKE_mesh_polygons_for_write(orgmesh);
if (mpoly) {
mp_example = *mpoly;
}
@@ -3273,9 +3260,9 @@ static Mesh *create_liquid_geometry(FluidDomainSettings *fds,
if (!me) {
return NULL;
}
- mverts = me->mvert;
- mpolys = me->mpoly;
- mloops = me->mloop;
+ mverts = BKE_mesh_vertices_for_write(me);
+ mpolys = BKE_mesh_polygons_for_write(me);
+ mloops = BKE_mesh_loops_for_write(me);
/* Get size (dimension) but considering scaling. */
copy_v3_v3(cell_size_scaled, fds->cell_size);
@@ -3409,9 +3396,9 @@ static Mesh *create_smoke_geometry(FluidDomainSettings *fds, Mesh *orgmesh, Obje
}
result = BKE_mesh_new_nomain(num_verts, 0, 0, num_faces * 4, num_faces);
- mverts = result->mvert;
- mpolys = result->mpoly;
- mloops = result->mloop;
+ mverts = BKE_mesh_vertices_for_write(result);
+ mpolys = BKE_mesh_polygons_for_write(result);
+ mloops = BKE_mesh_loops_for_write(result);
if (num_verts) {
/* Volume bounds. */