Displist: Reuse tesselated pos and nor for wireframes

This lower the memory usage and also fix a bug with metaballs normals/tris
winding being reversed.
This commit is contained in:
Clément Foucault 2018-12-14 02:52:50 +01:00
parent 18d0566013
commit 531e5ad49a
Notes: blender-bot 2023-02-14 01:57:12 +01:00
Referenced by commit d2c941def3, DRW: remove INDEX4 normal flipping
3 changed files with 105 additions and 94 deletions

View File

@ -899,6 +899,7 @@ void DRW_curve_batch_cache_create_requested(Object *ob)
DRW_vbo_request(cache->batch.curves, &cache->ordered.curves_pos);
}
if (DRW_batch_requested(cache->batch.wire_triangles, GPU_PRIM_TRIS)) {
DRW_vbo_request(cache->batch.wire_triangles, &cache->tess.pos_nor);
DRW_vbo_request(cache->batch.wire_triangles, &cache->tess.wireframe_data);
}

View File

@ -125,6 +125,51 @@ static void displist_indexbufbuilder_set(
}
}
static int displist_indexbufbuilder_tess_set(
setTriIndicesFn *set_tri_indices,
setTriIndicesFn *set_quad_tri_indices, /* meh, find a better solution. */
void *thunk, const DispList *dl, const int ofs)
{
int v_idx = ofs;
if (ELEM(dl->type, DL_INDEX3, DL_INDEX4, DL_SURF)) {
if (dl->type == DL_INDEX3) {
for (int i = 0; i < dl->parts; i++) {
set_tri_indices(thunk, v_idx + 0, v_idx + 1, v_idx + 2);
v_idx += 3;
}
}
else if (dl->type == DL_SURF) {
for (int a = 0; a < dl->parts; a++) {
if ((dl->flag & DL_CYCL_V) == 0 && a == dl->parts - 1) {
break;
}
int b = (dl->flag & DL_CYCL_U) ? 0 : 1;
for (; b < dl->nr; b++) {
set_quad_tri_indices(thunk, v_idx + 0, v_idx + 1, v_idx + 2);
set_quad_tri_indices(thunk, v_idx + 3, v_idx + 4, v_idx + 5);
v_idx += 6;
}
}
}
else {
BLI_assert(dl->type == DL_INDEX4);
const int *idx = dl->index;
for (int i = 0; i < dl->parts; i++, idx += 4) {
if (idx[2] != idx[3]) {
set_quad_tri_indices(thunk, v_idx + 0, v_idx + 1, v_idx + 2);
set_quad_tri_indices(thunk, v_idx + 3, v_idx + 4, v_idx + 5);
v_idx += 6;
}
else {
set_tri_indices(thunk, v_idx + 0, v_idx + 1, v_idx + 2);
v_idx += 3;
}
}
}
}
return v_idx;
}
void DRW_displist_vertbuf_create_pos_and_nor(ListBase *lb, GPUVertBuf *vbo)
{
static GPUVertFormat format = { 0 };
@ -199,42 +244,9 @@ void DRW_displist_indexbuf_create_triangles_tess_split_by_material(
/* calc each index buffer builder */
uint v_idx = 0;
for (const DispList *dl = lb->first; dl; dl = dl->next) {
if (ELEM(dl->type, DL_INDEX3, DL_INDEX4, DL_SURF)) {
GPUIndexBufBuilder *elem = &elb[dl->col];
if (dl->type == DL_INDEX3) {
for (int i = 0; i < dl->parts; i++) {
GPU_indexbuf_add_tri_verts(elem, v_idx + 0, v_idx + 1, v_idx + 2);
v_idx += 3;
}
}
else if (dl->type == DL_SURF) {
for (int a = 0; a < dl->parts; a++) {
if ((dl->flag & DL_CYCL_V) == 0 && a == dl->parts - 1) {
break;
}
int b = (dl->flag & DL_CYCL_U) ? 0 : 1;
for (; b < dl->nr; b++) {
/* TODO(fclem) reuse verts in a quad at least. */
GPU_indexbuf_add_tri_verts(elem, v_idx + 0, v_idx + 1, v_idx + 2);
GPU_indexbuf_add_tri_verts(elem, v_idx + 3, v_idx + 4, v_idx + 5);
v_idx += 6;
}
}
}
else {
BLI_assert(dl->type == DL_INDEX4);
const int *idx = dl->index;
for (int i = 0; i < dl->parts; i++, idx += 4) {
GPU_indexbuf_add_tri_verts(elem, v_idx + 0, v_idx + 1, v_idx + 2);
v_idx += 3;
if (idx[2] != idx[3]) {
GPU_indexbuf_add_tri_verts(elem, v_idx + 0, v_idx + 1, v_idx + 2);
v_idx += 3;
}
}
}
}
v_idx = displist_indexbufbuilder_tess_set((setTriIndicesFn *)GPU_indexbuf_add_tri_verts,
(setTriIndicesFn *)GPU_indexbuf_add_tri_verts,
&elb[dl->col], dl, v_idx);
}
/* build each indexbuf */
@ -244,7 +256,7 @@ void DRW_displist_indexbuf_create_triangles_tess_split_by_material(
}
typedef struct DRWDisplistWireThunk {
uint wd_id, pos_id, nor_id, vidx, ofs;
uint wd_id, ofs;
const DispList *dl;
GPUVertBuf *vbo;
} DRWDisplistWireThunk;
@ -252,44 +264,24 @@ typedef struct DRWDisplistWireThunk {
static void set_overlay_wires_tri_indices(void *thunk, uint v1, uint v2, uint v3)
{
DRWDisplistWireThunk *dwt = (DRWDisplistWireThunk *)thunk;
const DispList *dl = dwt->dl;
uint indices[3] = {v1, v2, v3};
const bool ndata_is_single = dl->type == DL_INDEX3;
for (int i = 0; i < 3; ++i) {
uint v = indices[i] - dwt->ofs;
/* TODO: Compute sharpness. For now, only tag real egdes. */
uchar sharpness = 0xFF;
short short_no[3];
const float(*verts)[3] = (float(*)[3])dl->verts;
const float(*nors)[3] = (float(*)[3])dl->nors;
normal_float_to_short_v3(short_no, nors[(ndata_is_single) ? 0 : v]);
GPU_vertbuf_attr_set(dwt->vbo, dwt->wd_id, dwt->vidx, &sharpness);
GPU_vertbuf_attr_set(dwt->vbo, dwt->pos_id, dwt->vidx, verts[v]);
GPU_vertbuf_attr_set(dwt->vbo, dwt->nor_id, dwt->vidx, short_no);
dwt->vidx++;
GPU_vertbuf_attr_set(dwt->vbo, dwt->wd_id, indices[i], &sharpness);
}
}
static void set_overlay_wires_quad_tri_indices(void *thunk, uint v1, uint v2, uint v3)
{
DRWDisplistWireThunk *dwt = (DRWDisplistWireThunk *)thunk;
const DispList *dl = dwt->dl;
uint indices[3] = {v1, v2, v3};
const bool ndata_is_single = dl->type == DL_INDEX3;
for (int i = 0; i < 3; ++i) {
uint v = indices[i] - dwt->ofs;
/* TODO: Compute sharpness. For now, only tag real egdes. */
uchar sharpness = (i == 0) ? 0x00 : 0xFF;
short short_no[3];
const float(*verts)[3] = (float(*)[3])dl->verts;
const float(*nors)[3] = (float(*)[3])dl->nors;
normal_float_to_short_v3(short_no, nors[(ndata_is_single) ? 0 : v]);
GPU_vertbuf_attr_set(dwt->vbo, dwt->wd_id, dwt->vidx, &sharpness);
GPU_vertbuf_attr_set(dwt->vbo, dwt->pos_id, dwt->vidx, verts[v]);
GPU_vertbuf_attr_set(dwt->vbo, dwt->nor_id, dwt->vidx, short_no);
dwt->vidx++;
GPU_vertbuf_attr_set(dwt->vbo, dwt->wd_id, indices[i], &sharpness);
}
}
@ -300,8 +292,6 @@ void DRW_displist_vertbuf_create_wireframe_data_tess(ListBase *lb, GPUVertBuf *v
static GPUVertFormat format = {0};
if (format.attr_len == 0) {
thunk.wd_id = GPU_vertformat_attr_add(&format, "wd", GPU_COMP_U8, 1, GPU_FETCH_INT_TO_FLOAT_UNIT);
thunk.pos_id = GPU_vertformat_attr_add(&format, "pos", GPU_COMP_F32, 3, GPU_FETCH_FLOAT);
thunk.nor_id = GPU_vertformat_attr_add(&format, "nor", GPU_COMP_I16, 3, GPU_FETCH_INT_TO_FLOAT_UNIT);
GPU_vertformat_triple_load(&format);
}
@ -311,21 +301,17 @@ void DRW_displist_vertbuf_create_wireframe_data_tess(ListBase *lb, GPUVertBuf *v
int vert_len = curve_render_surface_tri_len_get(lb) * 3;
GPU_vertbuf_data_alloc(thunk.vbo, vert_len);
thunk.vidx = 0;
thunk.ofs = 0;
for (const DispList *dl = lb->first; dl; dl = dl->next) {
thunk.dl = dl;
BKE_displist_normals_add(lb);
/* TODO consider non-manifold edges correctly. */
displist_indexbufbuilder_set(set_overlay_wires_tri_indices,
set_overlay_wires_quad_tri_indices,
&thunk, dl, thunk.ofs);
thunk.ofs += dl_vert_len(dl);
thunk.ofs = displist_indexbufbuilder_tess_set(set_overlay_wires_tri_indices,
set_overlay_wires_quad_tri_indices,
&thunk, dl, thunk.ofs);
}
if (thunk.vidx < vert_len) {
GPU_vertbuf_data_resize(thunk.vbo, thunk.vidx);
if (thunk.ofs < vert_len) {
GPU_vertbuf_data_resize(thunk.vbo, thunk.ofs);
}
}
@ -361,16 +347,28 @@ static void displist_vertbuf_attr_set_tri_pos_nor_uv(
GPUVertBufRaw *pos_step, GPUVertBufRaw *nor_step, GPUVertBufRaw *uv_step,
const float v1[3], const float v2[3], const float v3[3],
const float n1[3], const float n2[3], const float n3[3],
const float uv1[2], const float uv2[2], const float uv3[2])
const float uv1[2], const float uv2[2], const float uv3[2],
const bool invert_normal)
{
if (pos_step->size != 0) {
copy_v3_v3(GPU_vertbuf_raw_step(pos_step), v1);
copy_v3_v3(GPU_vertbuf_raw_step(pos_step), v2);
copy_v3_v3(GPU_vertbuf_raw_step(pos_step), v3);
*(GPUPackedNormal *)GPU_vertbuf_raw_step(nor_step) = GPU_normal_convert_i10_v3(n1);
*(GPUPackedNormal *)GPU_vertbuf_raw_step(nor_step) = GPU_normal_convert_i10_v3(n2);
*(GPUPackedNormal *)GPU_vertbuf_raw_step(nor_step) = GPU_normal_convert_i10_v3(n3);
if (invert_normal) {
float neg_n1[3], neg_n2[3], neg_n3[3];
negate_v3_v3(neg_n1, n1);
negate_v3_v3(neg_n2, n2);
negate_v3_v3(neg_n3, n3);
*(GPUPackedNormal *)GPU_vertbuf_raw_step(nor_step) = GPU_normal_convert_i10_v3(neg_n1);
*(GPUPackedNormal *)GPU_vertbuf_raw_step(nor_step) = GPU_normal_convert_i10_v3(neg_n2);
*(GPUPackedNormal *)GPU_vertbuf_raw_step(nor_step) = GPU_normal_convert_i10_v3(neg_n3);
}
else {
*(GPUPackedNormal *)GPU_vertbuf_raw_step(nor_step) = GPU_normal_convert_i10_v3(n1);
*(GPUPackedNormal *)GPU_vertbuf_raw_step(nor_step) = GPU_normal_convert_i10_v3(n2);
*(GPUPackedNormal *)GPU_vertbuf_raw_step(nor_step) = GPU_normal_convert_i10_v3(n3);
}
}
if (uv_step->size != 0) {
@ -391,6 +389,7 @@ void DRW_displist_vertbuf_create_pos_and_nor_and_uv_tess(
/* initialize vertex format */
attr_id.pos = GPU_vertformat_attr_add(&format_pos_nor, "pos", GPU_COMP_F32, 3, GPU_FETCH_FLOAT);
attr_id.nor = GPU_vertformat_attr_add(&format_pos_nor, "nor", GPU_COMP_I10, 3, GPU_FETCH_INT_TO_FLOAT_UNIT);
GPU_vertformat_triple_load(&format_pos_nor);
/* UVs are in [0..1] range. We can compress them. */
attr_id.uv = GPU_vertformat_attr_add(&format_uv, "u", GPU_COMP_I16, 2, GPU_FETCH_INT_TO_FLOAT_UNIT);
}
@ -429,15 +428,15 @@ void DRW_displist_vertbuf_create_pos_and_nor_and_uv_tess(
for (int i = 0; i < i_end; i++, idx += 3) {
if (vbo_uv) {
uv[0][0] = idx[0] / x_max;
uv[1][0] = idx[2] / x_max;
uv[2][0] = idx[1] / x_max;
uv[1][0] = idx[1] / x_max;
uv[2][0] = idx[2] / x_max;
}
displist_vertbuf_attr_set_tri_pos_nor_uv(
&pos_step, &nor_step, &uv_step,
verts[idx[0]], verts[idx[2]], verts[idx[1]],
dl->nors, dl->nors, dl->nors,
uv[0], uv[1], uv[2]);
uv[0], uv[2], uv[1], false);
}
}
else if (dl->type == DL_SURF) {
@ -474,15 +473,15 @@ void DRW_displist_vertbuf_create_pos_and_nor_and_uv_tess(
displist_vertbuf_attr_set_tri_pos_nor_uv(
&pos_step, &nor_step, &uv_step,
verts[quad[0]], verts[quad[1]], verts[quad[2]],
nors[quad[0]], nors[quad[1]], nors[quad[2]],
uv[0], uv[1], uv[2]);
verts[quad[2]], verts[quad[0]], verts[quad[1]],
nors[quad[2]], nors[quad[0]], nors[quad[1]],
uv[2], uv[0], uv[1], false);
displist_vertbuf_attr_set_tri_pos_nor_uv(
&pos_step, &nor_step, &uv_step,
verts[quad[0]], verts[quad[2]], verts[quad[3]],
nors[quad[0]], nors[quad[2]], nors[quad[3]],
uv[0], uv[2], uv[3]);
uv[0], uv[2], uv[3], false);
quad[2] = quad[1];
quad[1]++;
@ -500,27 +499,32 @@ void DRW_displist_vertbuf_create_pos_and_nor_and_uv_tess(
for (int i = 0; i < i_end; i++, idx += 4) {
displist_vertbuf_attr_set_tri_pos_nor_uv(
&pos_step, &nor_step, &uv_step,
verts[idx[0]], verts[idx[1]], verts[idx[2]],
nors[idx[0]], nors[idx[1]], nors[idx[2]],
uv[0], uv[1], uv[2]);
verts[idx[0]], verts[idx[2]], verts[idx[1]],
nors[idx[0]], nors[idx[2]], nors[idx[1]],
uv[0], uv[2], uv[1], true);
if (idx[2] != idx[3]) {
displist_vertbuf_attr_set_tri_pos_nor_uv(
&pos_step, &nor_step, &uv_step,
verts[idx[0]], verts[idx[2]], verts[idx[3]],
nors[idx[0]], nors[idx[2]], nors[idx[3]],
uv[0], uv[2], uv[3]);
verts[idx[2]], verts[idx[0]], verts[idx[3]],
nors[idx[2]], nors[idx[0]], nors[idx[3]],
uv[2], uv[0], uv[3], true);
}
}
}
}
}
#ifdef DEBUG
/* Resize and finish. */
if (pos_step.size != 0) {
BLI_assert(vbo_len_capacity == GPU_vertbuf_raw_used(&pos_step));
int vbo_len_used = GPU_vertbuf_raw_used(&pos_step);
if (vbo_len_used < vbo_len_capacity) {
GPU_vertbuf_data_resize(vbo_pos_nor, vbo_len_used);
}
}
if (uv_step.size != 0) {
BLI_assert(vbo_len_capacity == GPU_vertbuf_raw_used(&uv_step));
int vbo_len_used = GPU_vertbuf_raw_used(&uv_step);
if (vbo_len_used < vbo_len_capacity) {
GPU_vertbuf_data_resize(vbo_uv, vbo_len_used);
}
}
#endif
}

View File

@ -207,9 +207,15 @@ GPUBatch *DRW_metaball_batch_cache_get_wireframes_face(Object *ob)
if (cache->face_wire.batch == NULL) {
ListBase *lb = &ob->runtime.curve_cache->disp;
GPUVertBuf *vbo = MEM_callocN(sizeof(GPUVertBuf), __func__);
DRW_displist_vertbuf_create_wireframe_data_tess(lb, vbo);
cache->face_wire.batch = GPU_batch_create_ex(GPU_PRIM_TRIS, vbo, NULL, GPU_BATCH_OWNS_VBO);
GPUVertBuf *vbo_pos_nor = MEM_callocN(sizeof(GPUVertBuf), __func__);
GPUVertBuf *vbo_wireframe_data = MEM_callocN(sizeof(GPUVertBuf), __func__);
DRW_displist_vertbuf_create_pos_and_nor_and_uv_tess(lb, vbo_pos_nor, NULL);
DRW_displist_vertbuf_create_wireframe_data_tess(lb, vbo_wireframe_data);
cache->face_wire.batch = GPU_batch_create_ex(GPU_PRIM_TRIS, vbo_pos_nor, NULL, GPU_BATCH_OWNS_VBO);
GPU_batch_vertbuf_add_ex(cache->face_wire.batch, vbo_wireframe_data, true);
}
return cache->face_wire.batch;