ShuriZma
/
pcsx2
mirror of https://github.com/PCSX2/pcsx2.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

GS/HW: Cleanup texture shuffle shaders.

This commit is contained in:
lightningterror 2023-08-09 11:02:01 +02:00 committed by refractionpcsx2
parent 89688d67dd
commit 7fadb6ed05
4 changed files with 67 additions and 96 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
bin/resources/shaders/dx11/tfx.fx
View File

@ -942,20 +942,15 @@ PS_OUTPUT ps_main(PS_INPUT input)
uint4 denorm_c = uint4(C); uint4 denorm_c = uint4(C);
uint2 denorm_TA = uint2(float2(TA.xy) * 255.0f + 0.5f); uint2 denorm_TA = uint2(float2(TA.xy) * 255.0f + 0.5f);
// Special case for 32bit input and 16bit output, shuffle used by The Godfather. // Special case for 32bit input and 16bit output, shuffle used by The Godfather
if (PS_SHUFFLE_SAME) if (PS_SHUFFLE_SAME)
{ {
if (PS_READ_BA) if (PS_READ_BA)
{ C = (float4)(float((denorm_c.b & 0x7Fu) | (denorm_c.a & 0x80u)));
C.ga = (float2)(float((denorm_c.b & 0x7Fu) | (denorm_c.a & 0x80u)));
C.rb = C.ga;
}
else else
{
C.ga = C.rg; C.ga = C.rg;
C.rb = C.ga;
}
} }
// Copy of a 16bit source in to this target
else if (PS_READ16_SRC) else if (PS_READ16_SRC)
{ {
C.rb = (float2)float((denorm_c.r >> 3) | (((denorm_c.g >> 3) & 0x7u) << 5)); C.rb = (float2)float((denorm_c.r >> 3) | (((denorm_c.g >> 3) & 0x7u) << 5));
@ -964,28 +959,23 @@ PS_OUTPUT ps_main(PS_INPUT input)
else else
C.ga = (float2)float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.x & 0x80u)); C.ga = (float2)float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.x & 0x80u));
} }
// Write RB part. Mask will take care of the correct destination
else if (PS_READ_BA)
{
C.rb = C.bb;
if (denorm_c.a & 0x80u)
C.ga = (float2)(float((denorm_c.a & 0x7Fu) | (denorm_TA.y & 0x80u)));
else
C.ga = (float2)(float((denorm_c.a & 0x7Fu) | (denorm_TA.x & 0x80u)));
}
else else
{ {
// Mask will take care of the correct destination C.rb = C.rr;
if (PS_READ_BA) if (denorm_c.g & 0x80u)
C.rb = C.bb; C.ga = (float2)(float((denorm_c.g & 0x7Fu) | (denorm_TA.y & 0x80u)));
else
C.rb = C.rr;
if (PS_READ_BA)
{
if (denorm_c.a & 0x80u)
C.ga = (float2)(float((denorm_c.a & 0x7Fu) | (denorm_TA.y & 0x80u)));
else
C.ga = (float2)(float((denorm_c.a & 0x7Fu) | (denorm_TA.x & 0x80u)));
}
else else
{ C.ga = (float2)(float((denorm_c.g & 0x7Fu) | (denorm_TA.x & 0x80u)));
if (denorm_c.g & 0x80u)
C.ga = (float2)(float((denorm_c.g & 0x7Fu) | (denorm_TA.y & 0x80u)));
else
C.ga = (float2)(float((denorm_c.g & 0x7Fu) | (denorm_TA.x & 0x80u)));
}
} }
} }

24
bin/resources/shaders/opengl/tfx_fs.glsl
View File

@ -950,35 +950,30 @@ void ps_main()
#if PS_SHUFFLE #if PS_SHUFFLE
uvec4 denorm_c = uvec4(C); uvec4 denorm_c = uvec4(C);
uvec2 denorm_TA = uvec2(vec2(TA.xy) * 255.0f + 0.5f); uvec2 denorm_TA = uvec2(vec2(TA.xy) * 255.0f + 0.5f);
// Special case for 32bit input and 16bit output, shuffle used by The Godfather
#if PS_SHUFFLE_SAME #if PS_SHUFFLE_SAME
#if (PS_READ_BA) #if (PS_READ_BA)
C.ga = vec2(float((denorm_c.b & 0x7Fu) | (denorm_c.a & 0x80u))); C = vec4(float((denorm_c.b & 0x7Fu) | (denorm_c.a & 0x80u)));
C.rb = C.ga;
#else #else
C.ga = C.rg; C.ga = C.rg;
C.rb = C.ga;
#endif #endif
#else // Copy of a 16bit source in to this target
#if PS_READ16_SRC #elif PS_READ16_SRC
C.rb = vec2(float((denorm_c.r >> 3) | (((denorm_c.g >> 3) & 0x7u) << 5))); C.rb = vec2(float((denorm_c.r >> 3) | (((denorm_c.g >> 3) & 0x7u) << 5)));
if (bool(denorm_c.a & 0x80u)) if (bool(denorm_c.a & 0x80u))
C.ga = vec2(float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.y & 0x80u))); C.ga = vec2(float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.y & 0x80u)));
else else
C.ga = vec2(float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.x & 0x80u))); C.ga = vec2(float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.x & 0x80u)));
#else // Write RB part. Mask will take care of the correct destination
// Write RB part. Mask will take care of the correct destination #elif PS_READ_BA
#if PS_READ_BA
C.rb = C.bb; C.rb = C.bb;
#else
C.rb = C.rr;
#endif
// FIXME precompute my_TA & 0x80 // FIXME precompute my_TA & 0x80
// Write GA part. Mask will take care of the correct destination // Write GA part. Mask will take care of the correct destination
// Note: GLSL 4.50/GL_EXT_shader_integer_mix support a mix instruction to select a component\n" // Note: GLSL 4.50/GL_EXT_shader_integer_mix support a mix instruction to select a component\n"
// However Nvidia emulate it with an if (at least on kepler arch) ...\n" // However Nvidia emulate it with an if (at least on kepler arch) ...\n"
#if PS_READ_BA
// bit field operation requires GL4 HW. Could be nice to merge it with step/mix below // bit field operation requires GL4 HW. Could be nice to merge it with step/mix below
// uint my_ta = (bool(bitfieldExtract(denorm_c.a, 7, 1))) ? denorm_TA.y : denorm_TA.x; // uint my_ta = (bool(bitfieldExtract(denorm_c.a, 7, 1))) ? denorm_TA.y : denorm_TA.x;
// denorm_c.a = bitfieldInsert(denorm_c.a, bitfieldExtract(my_ta, 7, 1), 7, 1); // denorm_c.a = bitfieldInsert(denorm_c.a, bitfieldExtract(my_ta, 7, 1), 7, 1);
@ -990,6 +985,7 @@ void ps_main()
C.ga = vec2(float((denorm_c.a & 0x7Fu) | (denorm_TA.x & 0x80u))); C.ga = vec2(float((denorm_c.a & 0x7Fu) | (denorm_TA.x & 0x80u)));
#else #else
C.rb = C.rr;
if (bool(denorm_c.g & 0x80u)) if (bool(denorm_c.g & 0x80u))
C.ga = vec2(float((denorm_c.g & 0x7Fu) | (denorm_TA.y & 0x80u))); C.ga = vec2(float((denorm_c.g & 0x7Fu) | (denorm_TA.y & 0x80u)));
else else
@ -1001,9 +997,7 @@ void ps_main()
// float sel = step(128.0f, c.g); // float sel = step(128.0f, c.g);
// vec2 c_shuffle = vec2((denorm_c.gg & 0x7Fu) | (denorm_TA & 0x80u)); // vec2 c_shuffle = vec2((denorm_c.gg & 0x7Fu) | (denorm_TA & 0x80u));
// c.ga = mix(c_shuffle.xx, c_shuffle.yy, sel); // c.ga = mix(c_shuffle.xx, c_shuffle.yy, sel);
#endif // PS_READ_BA
#endif // READ16_SRC
#endif // PS_SHUFFLE_SAME #endif // PS_SHUFFLE_SAME
#endif // PS_SHUFFLE #endif // PS_SHUFFLE

55
bin/resources/shaders/vulkan/tfx.glsl
View File

@ -1200,44 +1200,35 @@ void main()
uvec4 denorm_c = uvec4(C); uvec4 denorm_c = uvec4(C);
uvec2 denorm_TA = uvec2(vec2(TA.xy) * 255.0f + 0.5f); uvec2 denorm_TA = uvec2(vec2(TA.xy) * 255.0f + 0.5f);
// Special case for 32bit input and 16bit output, shuffle used by The Godfather. // Special case for 32bit input and 16bit output, shuffle used by The Godfather
#if PS_SHUFFLE_SAME #if PS_SHUFFLE_SAME
#if (PS_READ_BA) #if (PS_READ_BA)
C.ga = vec2(float((denorm_c.b & 0x7Fu) | (denorm_c.a & 0x80u))); C = vec4(float((denorm_c.b & 0x7Fu) | (denorm_c.a & 0x80u)));
C.rb = C.ga;
#else #else
C.ga = C.rg; C.ga = C.rg;
C.rb = C.ga;
#endif #endif
// Copy of a 16bit source in to this target
#elif PS_READ16_SRC
C.rb = vec2(float((denorm_c.r >> 3) | (((denorm_c.g >> 3) & 0x7u) << 5)));
if ((denorm_c.a & 0x80u) != 0u)
C.ga = vec2(float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.y & 0x80u)));
else
C.ga = vec2(float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.x & 0x80u)));
// Write RB part. Mask will take care of the correct destination
#elif PS_READ_BA
C.rb = C.bb;
if ((denorm_c.a & 0x80u) != 0u)
C.ga = vec2(float((denorm_c.a & 0x7Fu) | (denorm_TA.y & 0x80u)));
else
C.ga = vec2(float((denorm_c.a & 0x7Fu) | (denorm_TA.x & 0x80u)));
#else #else
#if PS_READ16_SRC C.rb = C.rr;
C.rb = vec2(float((denorm_c.r >> 3) | (((denorm_c.g >> 3) & 0x7u) << 5))); if ((denorm_c.g & 0x80u) != 0u)
if ((denorm_c.a & 0x80u) != 0u) C.ga = vec2(float((denorm_c.g & 0x7Fu) | (denorm_TA.y & 0x80u)));
C.ga = vec2(float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.y & 0x80u))); else
else C.ga = vec2(float((denorm_c.g & 0x7Fu) | (denorm_TA.x & 0x80u)));
C.ga = vec2(float((denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.x & 0x80u))); #endif // PS_SHUFFLE_SAME
#else #endif // PS_SHUFFLE
// Mask will take care of the correct destination
#if PS_READ_BA
C.rb = C.bb;
#else
C.rb = C.rr;
#endif
#if PS_READ_BA
if ((denorm_c.a & 0x80u) != 0u)
C.ga = vec2(float((denorm_c.a & 0x7Fu) | (denorm_TA.y & 0x80u)));
else
C.ga = vec2(float((denorm_c.a & 0x7Fu) | (denorm_TA.x & 0x80u)));
#else
if ((denorm_c.g & 0x80u) != 0u)
C.ga = vec2(float((denorm_c.g & 0x7Fu) | (denorm_TA.y & 0x80u)));
else
C.ga = vec2(float((denorm_c.g & 0x7Fu) | (denorm_TA.x & 0x80u)));
#endif
#endif
#endif
#endif
// Must be done before alpha correction // Must be done before alpha correction

42
pcsx2/GS/Renderers/Metal/tfx.metal
View File

@ -1022,37 +1022,33 @@ struct PSMain
uint4 denorm_c = uint4(C); uint4 denorm_c = uint4(C);
uint2 denorm_TA = uint2(cb.ta * 255.5f); uint2 denorm_TA = uint2(cb.ta * 255.5f);
// Special case for 32bit input and 16bit output, shuffle used by The Godfather
if (PS_SHUFFLE_SAME) if (PS_SHUFFLE_SAME)
{ {
if (PS_READ_BA) if (PS_READ_BA)
{ C = (denorm_c.b & 0x7Fu) | (denorm_c.a & 0x80);
C.ga = (denorm_c.b & 0x7Fu) | (denorm_c.a & 0x80);
C.rb = C.ga;
}
else else
{
C.ga = C.rg; C.ga = C.rg;
C.rb = C.ga; }
} // Copy of a 16bit source in to this target
else if (PS_READ16_SRC)
{
C.rb = (denorm_c.r >> 3) | (((denorm_c.g >> 3) & 0x7u) << 5);
if (denorm_c.a & 0x80)
C.ga = (denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.y & 0x80);
else
C.ga = (denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.x & 0x80);
}
// Write RB part. Mask will take care of the correct destination
else if (PS_READ_BA)
{
C.rb = C.bb;
C.ga = (denorm_c.a & 0x7F) | (denorm_c.a & 0x80 ? denorm_TA.y & 0x80 : denorm_TA.x & 0x80);
} }
else else
{ {
if (PS_READ16_SRC) C.rb = C.rr;
{ C.ga = (denorm_c.g & 0x7F) | (denorm_c.g & 0x80 ? denorm_TA.y & 0x80 : denorm_TA.x & 0x80);
C.rb = (denorm_c.r >> 3) | (((denorm_c.g >> 3) & 0x7u) << 5);
if (denorm_c.a & 0x80)
C.ga = (denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.y & 0x80);
else
C.ga = (denorm_c.g >> 6) | ((denorm_c.b >> 3) << 2) | (denorm_TA.x & 0x80);
}
else
{
C.rb = PS_READ_BA ? C.bb : C.rr;
if (PS_READ_BA)
C.ga = (denorm_c.a & 0x7F) | (denorm_c.a & 0x80 ? denorm_TA.y & 0x80 : denorm_TA.x & 0x80);
else
C.ga = (denorm_c.g & 0x7F) | (denorm_c.g & 0x80 ? denorm_TA.y & 0x80 : denorm_TA.x & 0x80);
}
} }
} }