mirror of
https://github.com/Lime3DS/Lime3DS
synced 2024-12-26 09:02:44 -06:00
gl_rasterizer: Provide texture buffer offsets.
This allows us to move all data into one TBO.
This commit is contained in:
parent
f50e505a5a
commit
10fba63b94
4 changed files with 41 additions and 14 deletions
|
@ -76,6 +76,9 @@ RasterizerOpenGL::RasterizerOpenGL()
|
||||||
uniform_block_data.proctex_lut_dirty = true;
|
uniform_block_data.proctex_lut_dirty = true;
|
||||||
uniform_block_data.proctex_diff_lut_dirty = true;
|
uniform_block_data.proctex_diff_lut_dirty = true;
|
||||||
|
|
||||||
|
for (int i = 0; i < 24; i++)
|
||||||
|
uniform_block_data.data.lighting_lut_offset[i / 4][i % 4] = 256 * i;
|
||||||
|
|
||||||
glGetIntegerv(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT, &uniform_buffer_alignment);
|
glGetIntegerv(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT, &uniform_buffer_alignment);
|
||||||
uniform_size_aligned_vs =
|
uniform_size_aligned_vs =
|
||||||
Common::AlignUp<size_t>(sizeof(VSUniformData), uniform_buffer_alignment);
|
Common::AlignUp<size_t>(sizeof(VSUniformData), uniform_buffer_alignment);
|
||||||
|
|
|
@ -32,6 +32,7 @@ namespace GLShader {
|
||||||
static const std::string UniformBlockDef = R"(
|
static const std::string UniformBlockDef = R"(
|
||||||
#define NUM_TEV_STAGES 6
|
#define NUM_TEV_STAGES 6
|
||||||
#define NUM_LIGHTS 8
|
#define NUM_LIGHTS 8
|
||||||
|
#define NUM_LIGHTING_SAMPLERS 24
|
||||||
|
|
||||||
struct LightSrc {
|
struct LightSrc {
|
||||||
vec3 specular_0;
|
vec3 specular_0;
|
||||||
|
@ -55,6 +56,13 @@ layout (std140) uniform shader_data {
|
||||||
int scissor_y1;
|
int scissor_y1;
|
||||||
int scissor_x2;
|
int scissor_x2;
|
||||||
int scissor_y2;
|
int scissor_y2;
|
||||||
|
int fog_lut_offset;
|
||||||
|
int proctex_noise_lut_offset;
|
||||||
|
int proctex_color_map_offset;
|
||||||
|
int proctex_alpha_map_offset;
|
||||||
|
int proctex_lut_offset;
|
||||||
|
int proctex_diff_lut_offset;
|
||||||
|
ivec4 lighting_lut_offset[NUM_LIGHTING_SAMPLERS / 4];
|
||||||
vec3 fog_color;
|
vec3 fog_color;
|
||||||
vec2 proctex_noise_f;
|
vec2 proctex_noise_f;
|
||||||
vec2 proctex_noise_a;
|
vec2 proctex_noise_a;
|
||||||
|
@ -1017,7 +1025,7 @@ void AppendProcTexClamp(std::string& out, const std::string& var, ProcTexClamp m
|
||||||
}
|
}
|
||||||
|
|
||||||
void AppendProcTexCombineAndMap(std::string& out, ProcTexCombiner combiner,
|
void AppendProcTexCombineAndMap(std::string& out, ProcTexCombiner combiner,
|
||||||
const std::string& map_lut) {
|
const std::string& map_lut, const std::string& offset) {
|
||||||
std::string combined;
|
std::string combined;
|
||||||
switch (combiner) {
|
switch (combiner) {
|
||||||
case ProcTexCombiner::U:
|
case ProcTexCombiner::U:
|
||||||
|
@ -1055,7 +1063,7 @@ void AppendProcTexCombineAndMap(std::string& out, ProcTexCombiner combiner,
|
||||||
combined = "0.0";
|
combined = "0.0";
|
||||||
break;
|
break;
|
||||||
}
|
}
|
||||||
out += "ProcTexLookupLUT(" + map_lut + ", " + combined + ")";
|
out += "ProcTexLookupLUT(" + map_lut + ", " + offset + ", " + combined + ")";
|
||||||
}
|
}
|
||||||
|
|
||||||
void AppendProcTexSampler(std::string& out, const PicaFSConfig& config) {
|
void AppendProcTexSampler(std::string& out, const PicaFSConfig& config) {
|
||||||
|
@ -1064,12 +1072,12 @@ void AppendProcTexSampler(std::string& out, const PicaFSConfig& config) {
|
||||||
// coord=1.0 is lut[127]+lut_diff[127]. For other indices, the result is interpolated using
|
// coord=1.0 is lut[127]+lut_diff[127]. For other indices, the result is interpolated using
|
||||||
// value entries and difference entries.
|
// value entries and difference entries.
|
||||||
out += R"(
|
out += R"(
|
||||||
float ProcTexLookupLUT(samplerBuffer lut, float coord) {
|
float ProcTexLookupLUT(samplerBuffer lut, int offset, float coord) {
|
||||||
coord *= 128;
|
coord *= 128;
|
||||||
float index_i = clamp(floor(coord), 0.0, 127.0);
|
float index_i = clamp(floor(coord), 0.0, 127.0);
|
||||||
float index_f = coord - index_i; // fract() cannot be used here because 128.0 needs to be
|
float index_f = coord - index_i; // fract() cannot be used here because 128.0 needs to be
|
||||||
// extracted as index_i = 127.0 and index_f = 1.0
|
// extracted as index_i = 127.0 and index_f = 1.0
|
||||||
vec2 entry = texelFetch(lut, int(index_i)).rg;
|
vec2 entry = texelFetch(lut, int(index_i) + offset).rg;
|
||||||
return clamp(entry.r + entry.g * index_f, 0.0, 1.0);
|
return clamp(entry.r + entry.g * index_f, 0.0, 1.0);
|
||||||
}
|
}
|
||||||
)";
|
)";
|
||||||
|
@ -1105,8 +1113,8 @@ float ProcTexNoiseCoef(vec2 x) {
|
||||||
float g2 = ProcTexNoiseRand2D(point + vec2(0.0, 1.0)) * (frac.x + frac.y - 1.0);
|
float g2 = ProcTexNoiseRand2D(point + vec2(0.0, 1.0)) * (frac.x + frac.y - 1.0);
|
||||||
float g3 = ProcTexNoiseRand2D(point + vec2(1.0, 1.0)) * (frac.x + frac.y - 2.0);
|
float g3 = ProcTexNoiseRand2D(point + vec2(1.0, 1.0)) * (frac.x + frac.y - 2.0);
|
||||||
|
|
||||||
float x_noise = ProcTexLookupLUT(proctex_noise_lut, frac.x);
|
float x_noise = ProcTexLookupLUT(proctex_noise_lut, proctex_noise_lut_offset, frac.x);
|
||||||
float y_noise = ProcTexLookupLUT(proctex_noise_lut, frac.y);
|
float y_noise = ProcTexLookupLUT(proctex_noise_lut, proctex_noise_lut_offset, frac.y);
|
||||||
float x0 = mix(g0, g1, x_noise);
|
float x0 = mix(g0, g1, x_noise);
|
||||||
float x1 = mix(g2, g3, x_noise);
|
float x1 = mix(g2, g3, x_noise);
|
||||||
return mix(x0, x1, y_noise);
|
return mix(x0, x1, y_noise);
|
||||||
|
@ -1148,7 +1156,8 @@ float ProcTexNoiseCoef(vec2 x) {
|
||||||
|
|
||||||
// Combine and map
|
// Combine and map
|
||||||
out += "float lut_coord = ";
|
out += "float lut_coord = ";
|
||||||
AppendProcTexCombineAndMap(out, config.state.proctex.color_combiner, "proctex_color_map");
|
AppendProcTexCombineAndMap(out, config.state.proctex.color_combiner, "proctex_color_map",
|
||||||
|
"proctex_color_map_offset");
|
||||||
out += ";\n";
|
out += ";\n";
|
||||||
|
|
||||||
// Look up color
|
// Look up color
|
||||||
|
@ -1162,14 +1171,16 @@ float ProcTexNoiseCoef(vec2 x) {
|
||||||
out += "int lut_index_i = int(lut_coord) + " +
|
out += "int lut_index_i = int(lut_coord) + " +
|
||||||
std::to_string(config.state.proctex.lut_offset) + ";\n";
|
std::to_string(config.state.proctex.lut_offset) + ";\n";
|
||||||
out += "float lut_index_f = fract(lut_coord);\n";
|
out += "float lut_index_f = fract(lut_coord);\n";
|
||||||
out += "vec4 final_color = texelFetch(proctex_lut, lut_index_i) + lut_index_f * "
|
out += "vec4 final_color = texelFetch(proctex_lut, lut_index_i + proctex_lut_offset) + "
|
||||||
"texelFetch(proctex_diff_lut, lut_index_i);\n";
|
"lut_index_f * "
|
||||||
|
"texelFetch(proctex_diff_lut, lut_index_i + proctex_diff_lut_offset);\n";
|
||||||
break;
|
break;
|
||||||
case ProcTexFilter::Nearest:
|
case ProcTexFilter::Nearest:
|
||||||
case ProcTexFilter::NearestMipmapLinear:
|
case ProcTexFilter::NearestMipmapLinear:
|
||||||
case ProcTexFilter::NearestMipmapNearest:
|
case ProcTexFilter::NearestMipmapNearest:
|
||||||
out += "lut_coord += " + std::to_string(config.state.proctex.lut_offset) + ";\n";
|
out += "lut_coord += " + std::to_string(config.state.proctex.lut_offset) + ";\n";
|
||||||
out += "vec4 final_color = texelFetch(proctex_lut, int(round(lut_coord)));\n";
|
out += "vec4 final_color = texelFetch(proctex_lut, int(round(lut_coord)) + "
|
||||||
|
"proctex_lut_offset);\n";
|
||||||
break;
|
break;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -1177,7 +1188,8 @@ float ProcTexNoiseCoef(vec2 x) {
|
||||||
// Note: in separate alpha mode, the alpha channel skips the color LUT look up stage. It
|
// Note: in separate alpha mode, the alpha channel skips the color LUT look up stage. It
|
||||||
// uses the output of CombineAndMap directly instead.
|
// uses the output of CombineAndMap directly instead.
|
||||||
out += "float final_alpha = ";
|
out += "float final_alpha = ";
|
||||||
AppendProcTexCombineAndMap(out, config.state.proctex.alpha_combiner, "proctex_alpha_map");
|
AppendProcTexCombineAndMap(out, config.state.proctex.alpha_combiner, "proctex_alpha_map",
|
||||||
|
"proctex_alpha_map_offset");
|
||||||
out += ";\n";
|
out += ";\n";
|
||||||
out += "return vec4(final_color.xyz, final_alpha);\n}\n";
|
out += "return vec4(final_color.xyz, final_alpha);\n}\n";
|
||||||
} else {
|
} else {
|
||||||
|
@ -1238,7 +1250,7 @@ vec3 quaternion_rotate(vec4 q, vec3 v) {
|
||||||
}
|
}
|
||||||
|
|
||||||
float LookupLightingLUT(int lut_index, int index, float delta) {
|
float LookupLightingLUT(int lut_index, int index, float delta) {
|
||||||
vec2 entry = texelFetch(lighting_lut, lut_index * 256 + index).rg;
|
vec2 entry = texelFetch(lighting_lut, lighting_lut_offset[lut_index >> 2][lut_index & 3] + index).rg;
|
||||||
return entry.r + entry.g * delta;
|
return entry.r + entry.g * delta;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -1481,7 +1493,7 @@ vec4 secondary_fragment_color = vec4(0.0);
|
||||||
// Generate clamped fog factor from LUT for given fog index
|
// Generate clamped fog factor from LUT for given fog index
|
||||||
out += "float fog_i = clamp(floor(fog_index), 0.0, 127.0);\n";
|
out += "float fog_i = clamp(floor(fog_index), 0.0, 127.0);\n";
|
||||||
out += "float fog_f = fog_index - fog_i;\n";
|
out += "float fog_f = fog_index - fog_i;\n";
|
||||||
out += "vec2 fog_lut_entry = texelFetch(fog_lut, int(fog_i)).rg;\n";
|
out += "vec2 fog_lut_entry = texelFetch(fog_lut, int(fog_i) + fog_lut_offset).rg;\n";
|
||||||
out += "float fog_factor = fog_lut_entry.r + fog_lut_entry.g * fog_f;\n";
|
out += "float fog_factor = fog_lut_entry.r + fog_lut_entry.g * fog_f;\n";
|
||||||
out += "fog_factor = clamp(fog_factor, 0.0, 1.0);\n";
|
out += "fog_factor = clamp(fog_factor, 0.0, 1.0);\n";
|
||||||
|
|
||||||
|
|
|
@ -6,6 +6,7 @@
|
||||||
|
|
||||||
#include <memory>
|
#include <memory>
|
||||||
#include <glad/glad.h>
|
#include <glad/glad.h>
|
||||||
|
#include "video_core/regs_lighting.h"
|
||||||
#include "video_core/renderer_opengl/gl_resource_manager.h"
|
#include "video_core/renderer_opengl/gl_resource_manager.h"
|
||||||
#include "video_core/renderer_opengl/gl_shader_gen.h"
|
#include "video_core/renderer_opengl/gl_shader_gen.h"
|
||||||
#include "video_core/renderer_opengl/pica_to_gl.h"
|
#include "video_core/renderer_opengl/pica_to_gl.h"
|
||||||
|
@ -38,6 +39,13 @@ struct UniformData {
|
||||||
GLint scissor_y1;
|
GLint scissor_y1;
|
||||||
GLint scissor_x2;
|
GLint scissor_x2;
|
||||||
GLint scissor_y2;
|
GLint scissor_y2;
|
||||||
|
GLint fog_lut_offset;
|
||||||
|
GLint proctex_noise_lut_offset;
|
||||||
|
GLint proctex_color_map_offset;
|
||||||
|
GLint proctex_alpha_map_offset;
|
||||||
|
GLint proctex_lut_offset;
|
||||||
|
GLint proctex_diff_lut_offset;
|
||||||
|
alignas(16) GLivec4 lighting_lut_offset[Pica::LightingRegs::NumLightingSampler / 4];
|
||||||
alignas(16) GLvec3 fog_color;
|
alignas(16) GLvec3 fog_color;
|
||||||
alignas(8) GLvec2 proctex_noise_f;
|
alignas(8) GLvec2 proctex_noise_f;
|
||||||
alignas(8) GLvec2 proctex_noise_a;
|
alignas(8) GLvec2 proctex_noise_a;
|
||||||
|
@ -50,7 +58,7 @@ struct UniformData {
|
||||||
};
|
};
|
||||||
|
|
||||||
static_assert(
|
static_assert(
|
||||||
sizeof(UniformData) == 0x470,
|
sizeof(UniformData) == 0x4e0,
|
||||||
"The size of the UniformData structure has changed, update the structure in the shader");
|
"The size of the UniformData structure has changed, update the structure in the shader");
|
||||||
static_assert(sizeof(UniformData) < 16384,
|
static_assert(sizeof(UniformData) < 16384,
|
||||||
"UniformData structure must be less than 16kb as per the OpenGL spec");
|
"UniformData structure must be less than 16kb as per the OpenGL spec");
|
||||||
|
|
|
@ -23,6 +23,10 @@ using GLuvec2 = std::array<GLuint, 2>;
|
||||||
using GLuvec3 = std::array<GLuint, 3>;
|
using GLuvec3 = std::array<GLuint, 3>;
|
||||||
using GLuvec4 = std::array<GLuint, 4>;
|
using GLuvec4 = std::array<GLuint, 4>;
|
||||||
|
|
||||||
|
using GLivec2 = std::array<GLint, 2>;
|
||||||
|
using GLivec3 = std::array<GLint, 3>;
|
||||||
|
using GLivec4 = std::array<GLint, 4>;
|
||||||
|
|
||||||
namespace PicaToGL {
|
namespace PicaToGL {
|
||||||
|
|
||||||
inline GLenum TextureFilterMode(Pica::TexturingRegs::TextureConfig::TextureFilter mode) {
|
inline GLenum TextureFilterMode(Pica::TexturingRegs::TextureConfig::TextureFilter mode) {
|
||||||
|
|
Loading…
Reference in a new issue