obs-StreamFX/source/gs-vertexbuffer.cpp
Michael Fabian 'Xaymar' Dirks 1ebb0beac4 gs-vertexbuffer: Avoid relying on std for aligned types
Changes the GS::VertexBuffer storage to be one continuous buffer that is properly aligned and is also now used for GS::Vertex. This halves the necessary memory, removes reallocation cost and removes the copy necessary to get things onto the GPU.

Related: #9
2018-01-18 05:02:10 +01:00

200 lines
5.7 KiB
C++

/*
* Modern effects for a modern Streamer
* Copyright (C) 2017 Michael Fabian Dirks
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
*/
#include "gs-vertexbuffer.h"
#include "util-memory.h"
#include <stdexcept>
extern "C" {
#pragma warning( push )
#pragma warning( disable: 4201 )
#include <libobs/obs.h>
#pragma warning( pop )
}
#pragma region Constructor & Destructor
GS::VertexBuffer::VertexBuffer(uint32_t maximumVertices) {
if (maximumVertices > MAXIMUM_VERTICES) {
throw std::out_of_range("maximumVertices out of range");
}
// Assign limits.
m_capacity = maximumVertices;
m_layers = MAXIMUM_UVW_LAYERS;
// Allocate memory for data.
m_vertexbufferdata = gs_vbdata_create();
m_vertexbufferdata->num = m_capacity;
m_vertexbufferdata->points = m_positions = (vec3*)util::malloc_aligned(16, sizeof(vec3) * m_capacity);
m_vertexbufferdata->normals = m_normals = (vec3*)util::malloc_aligned(16, sizeof(vec3) * m_capacity);
m_vertexbufferdata->tangents = m_tangents = (vec3*)util::malloc_aligned(16, sizeof(vec3) * m_capacity);
m_vertexbufferdata->colors = m_colors = (uint32_t*)util::malloc_aligned(16, sizeof(uint32_t) * m_capacity);
m_vertexbufferdata->num_tex = m_layers;
m_vertexbufferdata->tvarray = m_layerdata = (gs_tvertarray*)util::malloc_aligned(16, sizeof(gs_tvertarray)* m_layers);
for (size_t n = 0; n < MAXIMUM_UVW_LAYERS; n++) {
m_layerdata[n].array = m_uvs[n] = (vec4*)util::malloc_aligned(16, sizeof(vec4) * m_capacity);
m_layerdata[n].width = 4;
}
// Allocate GPU
obs_enter_graphics();
m_vertexbuffer = gs_vertexbuffer_create(m_vertexbufferdata, GS_DYNAMIC);
std::memset(m_vertexbufferdata, 0, sizeof(gs_vb_data));
m_vertexbufferdata->num = m_capacity;
m_vertexbufferdata->num_tex = m_layers;
obs_leave_graphics();
if (!m_vertexbuffer) {
throw std::runtime_error("Failed to create vertex buffer.");
}
}
GS::VertexBuffer::VertexBuffer() : VertexBuffer(MAXIMUM_VERTICES) {}
GS::VertexBuffer::~VertexBuffer() {
if (m_positions) {
util::free_aligned(m_positions);
m_positions = nullptr;
}
if (m_normals) {
util::free_aligned(m_normals);
m_normals = nullptr;
}
if (m_tangents) {
util::free_aligned(m_tangents);
m_tangents = nullptr;
}
if (m_colors) {
util::free_aligned(m_colors);
m_colors = nullptr;
}
for (size_t n = 0; n < MAXIMUM_UVW_LAYERS; n++) {
if (m_uvs[n]) {
util::free_aligned(m_uvs[n]);
m_uvs[n] = nullptr;
}
}
if (m_vertexbufferdata) {
std::memset(m_vertexbufferdata, 0, sizeof(gs_vb_data));
if (!m_vertexbuffer) {
gs_vbdata_destroy(m_vertexbufferdata);
m_vertexbufferdata = nullptr;
}
}
if (m_vertexbuffer) {
obs_enter_graphics();
gs_vertexbuffer_destroy(m_vertexbuffer);
obs_leave_graphics();
m_vertexbuffer = nullptr;
}
}
#pragma endregion Constructor & Destructor
#pragma region Copy & Move Constructor
GS::VertexBuffer::VertexBuffer(VertexBuffer& other) : VertexBuffer(other.m_capacity) {
}
GS::VertexBuffer::VertexBuffer(gs_vertbuffer_t* vb) {
m_vertexbuffer = vb;
}
#pragma endregion Copy & Move Constructor
void GS::VertexBuffer::resize(size_t new_size) {
if (new_size > m_capacity) {
throw std::out_of_range("new_size out of range");
}
m_size = new_size;
}
size_t GS::VertexBuffer::size() {
return m_size;
}
bool GS::VertexBuffer::empty() {
return m_size == 0;
}
const GS::Vertex GS::VertexBuffer::at(size_t idx) {
if ((idx < 0) || (idx >= m_size)) {
throw std::out_of_range("idx out of range");
}
GS::Vertex vtx;
vtx.position = &m_positions[idx];
vtx.normal = &m_normals[idx];
vtx.tangent = &m_tangents[idx];
vtx.color = &m_colors[idx];
for (size_t n = 0; n < MAXIMUM_UVW_LAYERS; n++) {
vtx.uv[n] = &m_uvs[n][idx];
}
return vtx;
}
const GS::Vertex GS::VertexBuffer::operator[](const size_t pos) {
return at(pos);
}
void GS::VertexBuffer::set_uv_layers(uint32_t layers) {
m_layers = layers;
}
uint32_t GS::VertexBuffer::uv_layers() {
return m_layers;
}
gs_vertbuffer_t* GS::VertexBuffer::get(bool refreshGPU) {
if (!refreshGPU)
return m_vertexbuffer;
if (m_size > m_capacity)
throw std::out_of_range("size is larger than capacity");
// Update VertexBuffer data.
m_vertexbufferdata = gs_vertexbuffer_get_data(m_vertexbuffer);
std::memset(m_vertexbufferdata, 0, sizeof(gs_vb_data));
m_vertexbufferdata->num = m_capacity;
m_vertexbufferdata->points = m_positions;
m_vertexbufferdata->normals = m_normals;
m_vertexbufferdata->tangents = m_tangents;
m_vertexbufferdata->colors = m_colors;
m_vertexbufferdata->num_tex = m_layers;
m_vertexbufferdata->tvarray = m_layerdata;
for (size_t n = 0; n < MAXIMUM_UVW_LAYERS; n++) {
m_layerdata[n].array = m_uvs[n];
m_layerdata[n].width = 4;
}
// Update GPU
obs_enter_graphics();
gs_vertexbuffer_flush(m_vertexbuffer);
obs_leave_graphics();
// WORKAROUND: OBS Studio 20.x and below incorrectly deletes data that it doesn't own.
std::memset(m_vertexbufferdata, 0, sizeof(gs_vb_data));
m_vertexbufferdata->num = m_capacity;
m_vertexbufferdata->num_tex = m_layers;
for (uint32_t n = 0; n < m_layers; n++) {
m_layerdata[n].width = 4;
}
return m_vertexbuffer;
}
gs_vertbuffer_t* GS::VertexBuffer::get() {
return get(true);
}