blender
/
blender
129
397
Fork 572
Code Issues 6.2k Pull Requests 748 Projects 19 Wiki Activity

Armature: Speedup by removing unecessary check 

drw_batch_cache_generate_requested() is only needed when the geom needs to be
created.

Went from 37fps to 47fps with artificial testcase (lots of bones with one custom shape). Baseline 2.79 is 24fps.

Also fix the drw_shgroup_bone_custom_wire.
This commit is contained in:
Clément Foucault 2019-04-26 16:37:15 +02:00
parent d4827cfa81
commit 7a48e25003
1 changed files with 21 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
source/blender/draw/intern/draw_armature.c
View File

@ -350,9 +350,6 @@ static void drw_shgroup_bone_custom_solid(const float (*bone_mat)[4],
struct GPUBatch *ledges = DRW_cache_object_loose_edges_get(custom);
float final_bonemat[4][4];
/* XXXXXXX needs to be moved elsewhere. */
drw_batch_cache_generate_requested(custom);
if (surf || edges || ledges) {
mul_m4_m4m4(final_bonemat, g_data.ob->obmat, bone_mat);
}
@ -363,6 +360,9 @@ static void drw_shgroup_bone_custom_solid(const float (*bone_mat)[4],
DRWShadingGroup *shgrp_geom_solid = BLI_ghash_lookup(g_data.passes.custom_shapes, surf);
if (shgrp_geom_solid == NULL) {
/* TODO(fclem) needs to be moved elsewhere. */
drw_batch_cache_generate_requested(custom);
/* NOTE! g_data.transparent require a separate shading group if the
* object is transparent. This is done by passing a different ghash
* for transparent armature in pose mode. */
@ -377,6 +377,9 @@ static void drw_shgroup_bone_custom_solid(const float (*bone_mat)[4],
DRWShadingGroup *shgrp_geom_wire = BLI_ghash_lookup(g_data.passes.custom_shapes, edges);
if (shgrp_geom_wire == NULL) {
/* TODO(fclem) needs to be moved elsewhere. */
drw_batch_cache_generate_requested(custom);
shgrp_geom_wire = shgroup_instance_bone_shape_outline(
g_data.passes.bone_outline, edges, sh_cfg);
@ -389,6 +392,9 @@ static void drw_shgroup_bone_custom_solid(const float (*bone_mat)[4],
DRWShadingGroup *shgrp_geom_ledges = BLI_ghash_lookup(g_data.passes.custom_shapes, ledges);
if (shgrp_geom_ledges == NULL) {
/* TODO(fclem) needs to be moved elsewhere. */
drw_batch_cache_generate_requested(custom);
shgrp_geom_ledges = shgroup_instance_wire(g_data.passes.bone_wire, ledges);
BLI_ghash_insert(g_data.passes.custom_shapes, ledges, shgrp_geom_ledges);
@ -405,15 +411,20 @@ static void drw_shgroup_bone_custom_wire(const float (*bone_mat)[4],
/* grr, not re-using instances! */
struct GPUBatch *geom = DRW_cache_object_all_edges_get(custom);
/* XXXXXXX needs to be moved elsewhere. */
drw_batch_cache_generate_requested(custom);
if (geom) {
DRWShadingGroup *shgrp_geom_wire = shgroup_instance_wire(g_data.passes.bone_wire, geom);
float final_bonemat[4][4], final_color[4];
DRWShadingGroup *shgrp_geom_wire = BLI_ghash_lookup(g_data.passes.custom_shapes, geom);
if (shgrp_geom_wire == NULL) {
/* TODO(fclem) needs to be moved elsewhere. */
drw_batch_cache_generate_requested(custom);
shgrp_geom_wire = shgroup_instance_wire(g_data.passes.bone_wire, geom);
BLI_ghash_insert(g_data.passes.custom_shapes, geom, shgrp_geom_wire);
}
float final_color[4] = {color[0], color[1], color[2], 1.0f};
float final_bonemat[4][4];
mul_m4_m4m4(final_bonemat, g_data.ob->obmat, bone_mat);
copy_v3_v3(final_color, color);
final_color[3] = 1.0f; /* hack */
DRW_shgroup_call_dynamic_add(shgrp_geom_wire, final_bonemat, final_color);
}
}