mirror of
https://github.com/ZDoom/gzdoom.git
synced 2024-11-30 15:52:09 +00:00
Add light, blend and sampler functions
This commit is contained in:
parent
f9a7186550
commit
576fed5afc
4 changed files with 173 additions and 7 deletions
|
@ -154,8 +154,7 @@ void FixedFunction::CodegenDrawSpan()
|
||||||
SSAInt spot64 = ((xfrac >> (32 - 6 - 6))&(63 * 64)) + (yfrac >> (32 - 6));
|
SSAInt spot64 = ((xfrac >> (32 - 6 - 6))&(63 * 64)) + (yfrac >> (32 - 6));
|
||||||
//SSAInt spot = ((xfrac >> xshift) & xmask) + (yfrac >> yshift);
|
//SSAInt spot = ((xfrac >> xshift) & xmask) + (yfrac >> yshift);
|
||||||
|
|
||||||
//*loop.dest = LightBgra::shade_bgra(_source[loop.spot64()], _light, _shade_constants);
|
colors[i] = shade_bgra_simple(source[spot64 * 4].load_vec4ub(), light);
|
||||||
colors[i] = source[spot64 * 4].load_vec4ub() * light / 256;
|
|
||||||
|
|
||||||
stack_xfrac.store(xfrac + xstep);
|
stack_xfrac.store(xfrac + xstep);
|
||||||
stack_yfrac.store(yfrac + ystep);
|
stack_yfrac.store(yfrac + ystep);
|
||||||
|
@ -181,9 +180,7 @@ void FixedFunction::CodegenDrawSpan()
|
||||||
SSAInt spot64 = ((xfrac >> (32 - 6 - 6))&(63 * 64)) + (yfrac >> (32 - 6));
|
SSAInt spot64 = ((xfrac >> (32 - 6 - 6))&(63 * 64)) + (yfrac >> (32 - 6));
|
||||||
//SSAInt spot = ((xfrac >> xshift) & xmask) + (yfrac >> yshift);
|
//SSAInt spot = ((xfrac >> xshift) & xmask) + (yfrac >> yshift);
|
||||||
|
|
||||||
//*loop.dest = LightBgra::shade_bgra(_source[loop.spot64()], _light, _shade_constants);
|
SSAVec4i color = shade_bgra_simple(source[spot64 * 4].load_vec4ub(), light);
|
||||||
SSAVec4i color = source[spot64 * 4].load_vec4ub();
|
|
||||||
color = color * light / 256;
|
|
||||||
data[index * 4].store_vec4ub(color);
|
data[index * 4].store_vec4ub(color);
|
||||||
|
|
||||||
stack_index.store(index + 1);
|
stack_index.store(index + 1);
|
||||||
|
@ -200,6 +197,122 @@ void FixedFunction::CodegenDrawSpan()
|
||||||
mProgram.functionPassManager()->run(*function.func);
|
mProgram.functionPassManager()->run(*function.func);
|
||||||
}
|
}
|
||||||
|
|
||||||
|
SSAInt FixedFunction::calc_light_multiplier(SSAInt light)
|
||||||
|
{
|
||||||
|
return 256 - (light >> (FRACBITS - 8));
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::shade_pal_index_simple(SSAInt index, SSAInt light, SSAUBytePtr basecolors)
|
||||||
|
{
|
||||||
|
SSAVec4i color = basecolors[index * 4].load_vec4ub(); // = GPalette.BaseColors[index];
|
||||||
|
return shade_bgra_simple(color, light);
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::shade_pal_index_advanced(SSAInt index, SSAInt light, const SSAShadeConstants &constants, SSAUBytePtr basecolors)
|
||||||
|
{
|
||||||
|
SSAVec4i color = basecolors[index * 4].load_vec4ub(); // = GPalette.BaseColors[index];
|
||||||
|
return shade_bgra_advanced(color, light, constants);
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::shade_bgra_simple(SSAVec4i color, SSAInt light)
|
||||||
|
{
|
||||||
|
color = color * light / 256;
|
||||||
|
return color.insert(3, 255);
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::shade_bgra_advanced(SSAVec4i color, SSAInt light, const SSAShadeConstants &constants)
|
||||||
|
{
|
||||||
|
SSAInt blue = color[0];
|
||||||
|
SSAInt green = color[1];
|
||||||
|
SSAInt red = color[2];
|
||||||
|
SSAInt alpha = color[3];
|
||||||
|
|
||||||
|
SSAInt intensity = ((red * 77 + green * 143 + blue * 37) >> 8) * constants.desaturate;
|
||||||
|
|
||||||
|
SSAVec4i inv_light = 256 - light;
|
||||||
|
SSAVec4i inv_desaturate = 256 - constants.desaturate;
|
||||||
|
|
||||||
|
color = (color * inv_desaturate + intensity) / 256;
|
||||||
|
color = (constants.fade * inv_light + color * light) / 256;
|
||||||
|
color = (color * constants.light) / 256;
|
||||||
|
|
||||||
|
return color.insert(3, alpha);
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::blend_copy(SSAVec4i fg)
|
||||||
|
{
|
||||||
|
return fg;
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::blend_add(SSAVec4i fg, SSAVec4i bg, SSAInt srcalpha, SSAInt destalpha)
|
||||||
|
{
|
||||||
|
SSAVec4i color = (fg * srcalpha + bg * destalpha) / 256;
|
||||||
|
return color.insert(3, 255);
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::blend_sub(SSAVec4i fg, SSAVec4i bg, SSAInt srcalpha, SSAInt destalpha)
|
||||||
|
{
|
||||||
|
SSAVec4i color = (bg * destalpha - fg * srcalpha) / 256;
|
||||||
|
return color.insert(3, 255);
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::blend_revsub(SSAVec4i fg, SSAVec4i bg, SSAInt srcalpha, SSAInt destalpha)
|
||||||
|
{
|
||||||
|
SSAVec4i color = (fg * srcalpha - bg * destalpha) / 256;
|
||||||
|
return color.insert(3, 255);
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::blend_alpha_blend(SSAVec4i fg, SSAVec4i bg)
|
||||||
|
{
|
||||||
|
SSAInt alpha = fg[3];
|
||||||
|
alpha = alpha + (alpha >> 7); // // 255 -> 256
|
||||||
|
SSAInt inv_alpha = 256 - alpha;
|
||||||
|
SSAVec4i color = (fg * alpha + bg * inv_alpha) / 256;
|
||||||
|
return color.insert(3, 255);
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::sample_linear(SSAUBytePtr col0, SSAUBytePtr col1, SSAInt texturefracx, SSAInt texturefracy, SSAInt one, SSAInt height)
|
||||||
|
{
|
||||||
|
SSAInt frac_y0 = (texturefracy >> FRACBITS) * height;
|
||||||
|
SSAInt frac_y1 = ((texturefracy + one) >> FRACBITS) * height;
|
||||||
|
SSAInt y0 = frac_y0 >> FRACBITS;
|
||||||
|
SSAInt y1 = frac_y1 >> FRACBITS;
|
||||||
|
|
||||||
|
SSAVec4i p00 = col0[y0].load_vec4ub();
|
||||||
|
SSAVec4i p01 = col0[y1].load_vec4ub();
|
||||||
|
SSAVec4i p10 = col1[y0].load_vec4ub();
|
||||||
|
SSAVec4i p11 = col1[y1].load_vec4ub();
|
||||||
|
|
||||||
|
SSAInt inv_b = texturefracx;
|
||||||
|
SSAInt inv_a = (frac_y1 >> (FRACBITS - 4)) & 15;
|
||||||
|
SSAInt a = 16 - inv_a;
|
||||||
|
SSAInt b = 16 - inv_b;
|
||||||
|
|
||||||
|
return (p00 * (a * b) + p01 * (inv_a * b) + p10 * (a * inv_b) + p11 * (inv_a * inv_b) + 127) >> 8;
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i FixedFunction::sample_linear(SSAUBytePtr texture, SSAInt xfrac, SSAInt yfrac, SSAInt xbits, SSAInt ybits)
|
||||||
|
{
|
||||||
|
SSAInt xshift = (32 - xbits);
|
||||||
|
SSAInt yshift = (32 - ybits);
|
||||||
|
SSAInt xmask = (SSAInt(1) << xshift) - 1;
|
||||||
|
SSAInt ymask = (SSAInt(1) << yshift) - 1;
|
||||||
|
SSAInt x = xfrac >> xbits;
|
||||||
|
SSAInt y = yfrac >> ybits;
|
||||||
|
|
||||||
|
SSAVec4i p00 = texture[(y & ymask) + ((x & xmask) << yshift)].load_vec4ub();
|
||||||
|
SSAVec4i p01 = texture[((y + 1) & ymask) + ((x & xmask) << yshift)].load_vec4ub();
|
||||||
|
SSAVec4i p10 = texture[(y & ymask) + (((x + 1) & xmask) << yshift)].load_vec4ub();
|
||||||
|
SSAVec4i p11 = texture[((y + 1) & ymask) + (((x + 1) & xmask) << yshift)].load_vec4ub();
|
||||||
|
|
||||||
|
SSAInt inv_b = (xfrac >> (xbits - 4)) & 15;
|
||||||
|
SSAInt inv_a = (yfrac >> (ybits - 4)) & 15;
|
||||||
|
SSAInt a = 16 - inv_a;
|
||||||
|
SSAInt b = 16 - inv_b;
|
||||||
|
|
||||||
|
return (p00 * (a * b) + p01 * (inv_a * b) + p10 * (a * inv_b) + p11 * (inv_a * inv_b) + 127) >> 8;
|
||||||
|
}
|
||||||
|
|
||||||
llvm::Type *FixedFunction::GetRenderArgsStruct(llvm::LLVMContext &context)
|
llvm::Type *FixedFunction::GetRenderArgsStruct(llvm::LLVMContext &context)
|
||||||
{
|
{
|
||||||
std::vector<llvm::Type *> elements;
|
std::vector<llvm::Type *> elements;
|
||||||
|
|
|
@ -6,6 +6,7 @@
|
||||||
#include "r_compiler/ssa/ssa_vec8s.h"
|
#include "r_compiler/ssa/ssa_vec8s.h"
|
||||||
#include "r_compiler/ssa/ssa_vec16ub.h"
|
#include "r_compiler/ssa/ssa_vec16ub.h"
|
||||||
#include "r_compiler/ssa/ssa_int.h"
|
#include "r_compiler/ssa/ssa_int.h"
|
||||||
|
#include "r_compiler/ssa/ssa_int_ptr.h"
|
||||||
#include "r_compiler/ssa/ssa_short.h"
|
#include "r_compiler/ssa/ssa_short.h"
|
||||||
#include "r_compiler/ssa/ssa_ubyte_ptr.h"
|
#include "r_compiler/ssa/ssa_ubyte_ptr.h"
|
||||||
#include "r_compiler/ssa/ssa_vec4f_ptr.h"
|
#include "r_compiler/ssa/ssa_vec4f_ptr.h"
|
||||||
|
@ -57,8 +58,30 @@ struct RenderArgs
|
||||||
uint32_t light;
|
uint32_t light;
|
||||||
uint32_t srcalpha;
|
uint32_t srcalpha;
|
||||||
uint32_t destalpha;
|
uint32_t destalpha;
|
||||||
//ShadeConstants _shade_constants;
|
|
||||||
//int32_t nearest_filter;
|
uint16_t light_alpha;
|
||||||
|
uint16_t light_red;
|
||||||
|
uint16_t light_green;
|
||||||
|
uint16_t light_blue;
|
||||||
|
uint16_t fade_alpha;
|
||||||
|
uint16_t fade_red;
|
||||||
|
uint16_t fade_green;
|
||||||
|
uint16_t fade_blue;
|
||||||
|
uint16_t desaturate;
|
||||||
|
uint32_t flags;
|
||||||
|
enum Flags
|
||||||
|
{
|
||||||
|
simple_shade = 1,
|
||||||
|
nearest_filter = 2
|
||||||
|
};
|
||||||
|
};
|
||||||
|
|
||||||
|
class SSAShadeConstants
|
||||||
|
{
|
||||||
|
public:
|
||||||
|
SSAVec4i light;
|
||||||
|
SSAVec4i fade;
|
||||||
|
SSAInt desaturate;
|
||||||
};
|
};
|
||||||
|
|
||||||
class FixedFunction
|
class FixedFunction
|
||||||
|
@ -71,6 +94,24 @@ public:
|
||||||
private:
|
private:
|
||||||
void CodegenDrawSpan();
|
void CodegenDrawSpan();
|
||||||
|
|
||||||
|
// LightBgra
|
||||||
|
SSAInt calc_light_multiplier(SSAInt light);
|
||||||
|
SSAVec4i shade_pal_index_simple(SSAInt index, SSAInt light, SSAUBytePtr basecolors);
|
||||||
|
SSAVec4i shade_pal_index_advanced(SSAInt index, SSAInt light, const SSAShadeConstants &constants, SSAUBytePtr basecolors);
|
||||||
|
SSAVec4i shade_bgra_simple(SSAVec4i color, SSAInt light);
|
||||||
|
SSAVec4i shade_bgra_advanced(SSAVec4i color, SSAInt light, const SSAShadeConstants &constants);
|
||||||
|
|
||||||
|
// BlendBgra
|
||||||
|
SSAVec4i blend_copy(SSAVec4i fg);
|
||||||
|
SSAVec4i blend_add(SSAVec4i fg, SSAVec4i bg, SSAInt srcalpha, SSAInt destalpha);
|
||||||
|
SSAVec4i blend_sub(SSAVec4i fg, SSAVec4i bg, SSAInt srcalpha, SSAInt destalpha);
|
||||||
|
SSAVec4i blend_revsub(SSAVec4i fg, SSAVec4i bg, SSAInt srcalpha, SSAInt destalpha);
|
||||||
|
SSAVec4i blend_alpha_blend(SSAVec4i fg, SSAVec4i bg);
|
||||||
|
|
||||||
|
// SampleBgra
|
||||||
|
SSAVec4i sample_linear(SSAUBytePtr col0, SSAUBytePtr col1, SSAInt texturefracx, SSAInt texturefracy, SSAInt one, SSAInt height);
|
||||||
|
SSAVec4i sample_linear(SSAUBytePtr texture, SSAInt xfrac, SSAInt yfrac, SSAInt xbits, SSAInt ybits);
|
||||||
|
|
||||||
static llvm::Type *GetRenderArgsStruct(llvm::LLVMContext &context);
|
static llvm::Type *GetRenderArgsStruct(llvm::LLVMContext &context);
|
||||||
|
|
||||||
RenderProgram mProgram;
|
RenderProgram mProgram;
|
||||||
|
|
|
@ -60,6 +60,16 @@ SSAInt SSAVec4i::operator[](SSAInt index)
|
||||||
return SSAInt::from_llvm(SSAScope::builder().CreateExtractElement(v, index.v, SSAScope::hint()));
|
return SSAInt::from_llvm(SSAScope::builder().CreateExtractElement(v, index.v, SSAScope::hint()));
|
||||||
}
|
}
|
||||||
|
|
||||||
|
SSAVec4i SSAVec4i::insert(SSAInt index, SSAInt value)
|
||||||
|
{
|
||||||
|
return SSAVec4i::from_llvm(SSAScope::builder().CreateInsertElement(v, value.v, index.v, SSAScope::hint()));
|
||||||
|
}
|
||||||
|
|
||||||
|
SSAVec4i SSAVec4i::insert(int index, SSAInt value)
|
||||||
|
{
|
||||||
|
return SSAVec4i::from_llvm(SSAScope::builder().CreateInsertElement(v, value.v, index, SSAScope::hint()));
|
||||||
|
}
|
||||||
|
|
||||||
llvm::Type *SSAVec4i::llvm_type()
|
llvm::Type *SSAVec4i::llvm_type()
|
||||||
{
|
{
|
||||||
return llvm::VectorType::get(llvm::Type::getInt32Ty(SSAScope::context()), 4);
|
return llvm::VectorType::get(llvm::Type::getInt32Ty(SSAScope::context()), 4);
|
||||||
|
|
|
@ -19,6 +19,8 @@ public:
|
||||||
explicit SSAVec4i(llvm::Value *v);
|
explicit SSAVec4i(llvm::Value *v);
|
||||||
SSAVec4i(SSAVec4f f32);
|
SSAVec4i(SSAVec4f f32);
|
||||||
SSAInt operator[](SSAInt index);
|
SSAInt operator[](SSAInt index);
|
||||||
|
SSAVec4i insert(SSAInt index, SSAInt value);
|
||||||
|
SSAVec4i insert(int index, SSAInt value);
|
||||||
static SSAVec4i bitcast(SSAVec4f f32);
|
static SSAVec4i bitcast(SSAVec4f f32);
|
||||||
static SSAVec4i bitcast(SSAVec8s i16);
|
static SSAVec4i bitcast(SSAVec8s i16);
|
||||||
static SSAVec4i shuffle(const SSAVec4i &f0, int index0, int index1, int index2, int index3);
|
static SSAVec4i shuffle(const SSAVec4i &f0, int index0, int index1, int index2, int index3);
|
||||||
|
|
Loading…
Reference in a new issue