raphiaura/KytyPS5
1
0
Fork 0
mirror of https://github.com/KytyPS5/KytyPS5.git synced 2026-10-21 03:16:54 +02:00
Code Issues Packages Projects Releases Wiki Activity

shader: pack v_cvt_pkrtz_f16_f32 with the native half conversion (GTA V) (#1520)

Browse Source 
* shader: pack v_cvt_pkrtz_f16_f32 with the native half conversion

The round-toward-zero pack was emulated with integer operations: about 20 to 30 operations
per half. GTA V uses it in 613 of its 656 pixel shaders, about 16 times in each G-buffer
pixel shader.

The pair is now packed with PackHalf2x16, unpacked again, and each half whose magnitude
exceeds its input's takes one unit off. Half signs and magnitudes are separate, so that is
the next value toward zero, whatever rounding the native conversion uses as long as it
returns one of the two neighbours. Same bits as before for every finite input and infinity;
a NaN stays a NaN. Checked against the old code on 2 million values on the CPU, and on an
Apple M5 GPU through Metal (400 000 values, subnormals included, safe and fast math).

GTA V on an Apple M5 (MoltenVK), 1080p, fixed script camera, three places, A B B A between
launches: -1.7 to -2.8 ms per frame (+4 to +6 %).

Scope: generic
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* shader: remove half-pack commentary and unused shift helper

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: nmzik <Nmzik@mail.ru>
This commit is contained in:
AN0DIN
2026-10-16 02:56:35 +02:00
committed by GitHub
parent 13e4c315b6
commit 83db2efec3
3 changed files with 17 additions and 76 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/graphics/shader/recompiler/backend/spirv/spirvEmitterAlu.cpp
+17 -4
View File
@@ -210,10 +210,23 @@ uint32_t EmitCompositeExtractU32x2(EmitterState& state, uint32_t arg0, IR::Value
}
uint32_t EmitPackFloat2x16Rtz(EmitterState& state, uint32_t arg0, uint32_t arg1) {
const auto low = EmitF32ToF16RtzBits(state, arg0);
const auto high = Binary(state, spv::OpShiftLeftLogical, TypeU32(state),
EmitF32ToF16RtzBits(state, arg1), ConstantU32(state, 16));
return Binary(state, spv::OpBitwiseOr, TypeU32(state), low, high);
const auto pair = state.builder.AllocateId();
state.builder.AddFunction(spv::OpCompositeConstruct, TypeF32Vector(state, 2), pair, arg0, arg1);
const auto packed = EmitPackHalf2x16(state, pair);
const auto back = EmitExt(state, TypeF32Vector(state, 2), GLSLstd450UnpackHalf2x16, {packed});
auto result = packed;
for (uint32_t i = 0; i < 2; i++) {
const auto rounded = state.builder.AllocateId();
state.builder.AddFunction(spv::OpCompositeExtract, TypeF32(state), rounded, back, i);
const auto above =
Binary(state, spv::OpFOrdGreaterThan, TypeBool(state),
EmitExt(state, TypeF32(state), GLSLstd450FAbs, {rounded}),
EmitExt(state, TypeF32(state), GLSLstd450FAbs, {i == 0 ? arg0 : arg1}));
result = Binary(state, spv::OpISub, TypeU32(state), result,
Select(state, TypeU32(state), above,
ConstantU32(state, i == 0 ? 1u : 0x10000u), ConstantU32(state, 0)));
}
return result;
}
uint32_t EmitFPSaturate32(EmitterState& state, uint32_t arg0) {