Clay: Performance: Disable AO codepath if not necessary.

This optimisation only works if no material in the scene require the AO pass.
For this either set the AO distance to 0 or both Cavity and Edges factors to 0.

This double the performance of scenes with very high triangle count.

source/blender/draw/engines/clay/clay_engine.c

@@ -183,6 +183,7 @@ typedef struct CLAY_PrivateData {
 	DRWShadingGroup *depth_shgrp_cull;
 	DRWShadingGroup *depth_shgrp_cull_select;
 	DRWShadingGroup *depth_shgrp_cull_active;
+	bool enable_ao;
 } CLAY_PrivateData; /* Transient data */
 
 /* Functions */
@@ -607,7 +608,7 @@ static int hair_mat_in_ubo(CLAY_Storage *storage, const CLAY_HAIR_UBO_Material *
 	return id;
 }
 
-static void ubo_mat_from_object(Object *ob,  CLAY_UBO_Material *r_ubo)
+static void ubo_mat_from_object(Object *ob, CLAY_UBO_Material *r_ubo, bool *r_needs_ao)
 {
 	IDProperty *props = BKE_layer_collection_engine_evaluated_get(ob, COLLECTION_MODE_NONE, RE_engine_id_BLENDER_CLAY);
 
@@ -622,6 +623,12 @@ static void ubo_mat_from_object(Object *ob,  CLAY_UBO_Material *r_ubo)
 	float ssao_attenuation = BKE_collection_engine_property_value_get_float(props, "ssao_attenuation");
 	int matcap_icon = BKE_collection_engine_property_value_get_int(props, "matcap_icon");
 
+	if (((ssao_factor_cavity > 0.0) || (ssao_factor_edge > 0.0)) &&
+	    (ssao_distance > 0.0))
+	{
+		*r_needs_ao = true;
+	}
+
 	memset(r_ubo, 0x0, sizeof(*r_ubo));
 
 	r_ubo->matcap_rot[0] = cosf(matcap_rot * 3.14159f * 2.0f);
@@ -667,7 +674,7 @@ static DRWShadingGroup *CLAY_object_shgrp_get(
 	DRWShadingGroup **shgrps = use_flat ? stl->storage->shgrps_flat : stl->storage->shgrps;
 	CLAY_UBO_Material mat_ubo_test;
 
-	ubo_mat_from_object(ob, &mat_ubo_test);
+	ubo_mat_from_object(ob, &mat_ubo_test, &stl->g_data->enable_ao);
 
 	int id = mat_in_ubo(stl->storage, &mat_ubo_test);
 
@@ -712,6 +719,9 @@ static void clay_cache_init(void *vedata)
 		stl->g_data = MEM_mallocN(sizeof(*stl->g_data), __func__);
 	}
 
+	/* Disable AO unless a material needs it. */
+	stl->g_data->enable_ao = false;
+
 	/* Depth Pass */
 	{
 		psl->depth_pass = DRW_pass_create("Depth Pass", DRW_STATE_WRITE_DEPTH | DRW_STATE_DEPTH_LESS);
@@ -851,18 +861,25 @@ static void clay_cache_finish(void *vedata)
 
 static void clay_draw_scene(void *vedata)
 {
-
+	CLAY_StorageList *stl = ((CLAY_Data *)vedata)->stl;
 	CLAY_PassList *psl = ((CLAY_Data *)vedata)->psl;
 	CLAY_FramebufferList *fbl = ((CLAY_Data *)vedata)->fbl;
 	DefaultFramebufferList *dfbl = DRW_viewport_framebuffer_list_get();
 
 	/* Pass 1 : Depth pre-pass */
-	DRW_draw_pass(psl->depth_pass);
-	DRW_draw_pass(psl->depth_pass_cull);
+	if (stl->g_data->enable_ao) {
+		DRW_draw_pass(psl->depth_pass);
+		DRW_draw_pass(psl->depth_pass_cull);
+	}
+	else {
+		DRWState state = DRW_STATE_WRITE_COLOR | DRW_STATE_WRITE_DEPTH | DRW_STATE_DEPTH_LESS;
+		DRW_pass_state_set(psl->clay_pass, state);
+		DRW_pass_state_set(psl->clay_pass_flat, state);
+	}
 
 	/* Pass 2 : Duplicate depth */
 	/* Unless we go for deferred shading we need this to avoid manual depth test and artifacts */
-	if (DRW_state_is_fbo()) {
+	if (DRW_state_is_fbo() && stl->g_data->enable_ao) {
 		/* attach temp textures */
 		DRW_framebuffer_texture_attach(fbl->dupli_depth, e_data.depth_dup, 0, 0);
 

source/blender/draw/intern/DRW_render.h

@@ -358,6 +358,7 @@ void DRW_shgroup_uniform_mat4(DRWShadingGroup *shgroup, const char *name, const
 
 /* Passes */
 DRWPass *DRW_pass_create(const char *name, DRWState state);
+void DRW_pass_state_set(DRWPass *pass, DRWState state);
 void DRW_pass_foreach_shgroup(DRWPass *pass, void (*callback)(void *userData, DRWShadingGroup *shgrp), void *userData);
 void DRW_pass_sort_shgroup_z(DRWPass *pass);
 

source/blender/draw/intern/draw_manager.c

@@ -1290,6 +1290,11 @@ DRWPass *DRW_pass_create(const char *name, DRWState state)
 	return pass;
 }
 
+void DRW_pass_state_set(DRWPass *pass, DRWState state)
+{
+	pass->state = state;
+}
+
 void DRW_pass_free(DRWPass *pass)
 {
 	for (DRWShadingGroup *shgroup = pass->shgroups; shgroup; shgroup = shgroup->next) {