This repository has been archived on 2023-10-09. You can view files and clone it. You cannot open issues or pull requests or push a commit.
Files
blender-archive/source/blender/draw/intern/shaders/draw_command_generate_comp.glsl
Miguel Pozo 59b9bb0849 Draw: Custom IDs
This pull request adds a new tipe of resource handles (thin handles).
These are intended for cases where a resource buffer with more than one
entry for each object is needed (for example, one entry per material
slot).
While it's already possible to have multiple regular handles for the
same object, they have a non-trivial overhead in terms of uploaded
data (matrix, bounds, object info) and computation (visibility
culling).
Thin handles store an indirection buffer pointing to their "parent"
regular handle, therefore multiple thin handles can share the same
per-object data and visibility culling computation.

Thin handles can only be used in their own Pass type (PassMainThin),
so passes that don't need them don't have to pay the overhead.

This pull request also includes the update of the Workbench Next
pre-pass to use PassMainThin, which is the main reason for the
implementation of this feature.

The main change from the previous PR is that the thin handles are now
stored directly in the main resource_id_buf, to avoid wasting an extra
 bind slot.

Pull Request #105261
2023-03-01 21:42:25 +01:00

137 lines
4.4 KiB
GLSL

/**
* Convert DrawPrototype into draw commands.
*/
#pragma BLENDER_REQUIRE(common_math_lib.glsl)
#define atomicAddAndGet(dst, val) (atomicAdd(dst, val) + val)
/* This is only called by the last thread executed over the group's prototype draws. */
void write_draw_call(DrawGroup group, uint group_id)
{
DrawCommand cmd;
cmd.vertex_len = group.vertex_len;
cmd.vertex_first = group.vertex_first;
bool indexed_draw = group.base_index != -1;
if (indexed_draw) {
cmd.base_index = group.base_index;
cmd.instance_first_indexed = group.start;
}
else {
cmd._instance_first_array = group.start;
}
/* Back-facing command. */
cmd.instance_len = group_buf[group_id].back_facing_counter;
command_buf[group_id * 2 + 0] = cmd;
/* Front-facing command. */
uint front_facing_start = group.start + (group.len - group.front_facing_len);
if (indexed_draw) {
cmd.instance_first_indexed = front_facing_start;
}
else {
cmd._instance_first_array = front_facing_start;
}
cmd.instance_len = group_buf[group_id].front_facing_counter;
command_buf[group_id * 2 + 1] = cmd;
/* Reset the counters for a next command gen dispatch. Avoids resending the whole data just
* for this purpose. Only the last thread will execute this so it is thread-safe. */
group_buf[group_id].front_facing_counter = 0u;
group_buf[group_id].back_facing_counter = 0u;
group_buf[group_id].total_counter = 0u;
}
void main()
{
uint proto_id = gl_GlobalInvocationID.x;
if (proto_id >= prototype_len) {
return;
}
DrawPrototype proto = prototype_buf[proto_id];
uint group_id = proto.group_id;
bool is_inverted = (proto.resource_handle & 0x80000000u) != 0;
uint resource_index = (proto.resource_handle & 0x7FFFFFFFu);
/* Visibility test result. */
uint visible_instance_len = 0;
if (visibility_word_per_draw > 0) {
uint visibility_word = resource_index * visibility_word_per_draw;
for (uint i = 0; i < visibility_word_per_draw; i++, visibility_word++) {
/* NOTE: This assumes `proto.instance_len` is 1. */
/* TODO: Assert. */
visible_instance_len += bitCount(visibility_buf[visibility_word]);
}
}
else {
if ((visibility_buf[resource_index / 32u] & (1u << (resource_index % 32u))) != 0) {
visible_instance_len = proto.instance_len;
}
}
bool is_visible = visible_instance_len > 0;
DrawGroup group = group_buf[group_id];
if (!is_visible) {
/* Skip the draw but still count towards the completion. */
if (atomicAddAndGet(group_buf[group_id].total_counter, proto.instance_len) == group.len) {
write_draw_call(group, group_id);
}
return;
}
uint back_facing_len = group.len - group.front_facing_len;
uint front_facing_len = group.front_facing_len;
uint dst_index = group.start;
if (is_inverted) {
uint offset = atomicAdd(group_buf[group_id].back_facing_counter, visible_instance_len);
dst_index += offset;
if (atomicAddAndGet(group_buf[group_id].total_counter, proto.instance_len) == group.len) {
write_draw_call(group, group_id);
}
}
else {
uint offset = atomicAdd(group_buf[group_id].front_facing_counter, visible_instance_len);
dst_index += back_facing_len + offset;
if (atomicAddAndGet(group_buf[group_id].total_counter, proto.instance_len) == group.len) {
write_draw_call(group, group_id);
}
}
/* Fill resource_id buffer for each instance of this draw. */
if (visibility_word_per_draw > 0) {
uint visibility_word = resource_index * visibility_word_per_draw;
for (uint i = 0; i < visibility_word_per_draw; i++, visibility_word++) {
uint word = visibility_buf[visibility_word];
uint view_index = i * 32u;
while (word != 0u) {
if ((word & 1u) != 0u) {
if (use_custom_ids) {
resource_id_buf[dst_index * 2] = view_index | (resource_index << view_shift);
resource_id_buf[dst_index * 2 + 1] = proto.custom_id;
}
else {
resource_id_buf[dst_index] = view_index | (resource_index << view_shift);
}
dst_index++;
}
view_index++;
word >>= 1u;
}
}
}
else {
for (uint i = dst_index; i < dst_index + visible_instance_len; i++) {
if (use_custom_ids) {
resource_id_buf[i * 2] = resource_index;
resource_id_buf[i * 2 + 1] = proto.custom_id;
}
else {
resource_id_buf[i] = resource_index;
}
}
}
}