mirrors/shadPS4
1
0
Fork 0
mirror of https://github.com/shadps4-emu/shadPS4.git synced 2025-08-03 16:02:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix packing

Browse Source 
pack is for floats, we have uints here
This commit is contained in:
martin 2024-09-04 15:57:52 -05:00
parent 12af52348e
commit 43892c8679
2 changed files with 14 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
src/shader_recompiler/frontend/translate/translate.h
View File

@ -233,7 +233,7 @@ private:
[[nodiscard]] T GetSrc64(const InstOperand& operand); [[nodiscard]] T GetSrc64(const InstOperand& operand);
void SetDst(const InstOperand& operand, const IR::U32F32& value); void SetDst(const InstOperand& operand, const IR::U32F32& value);
void SetDst64(const InstOperand& operand, const IR::U64F64& value_raw); void SetDst64(const InstOperand& operand, const IR::U64F64& value_raw);
IR::U16 Convert_F32_to_U16_Normalized(const IR::F32& src); IR::U32 Convert_F32_to_U16_Normalized(const IR::F32& src);
void LogMissingOpcode(const GcnInst& inst); void LogMissingOpcode(const GcnInst& inst);
private: private:

25
src/shader_recompiler/frontend/translate/vector_alu.cpp
View File

@ -344,25 +344,26 @@ void Translator::V_CVT_PKRTZ_F16_F32(const GcnInst& inst) {
ir.SetVectorReg(dst_reg, ir.PackHalf2x16(vec_f32)); ir.SetVectorReg(dst_reg, ir.PackHalf2x16(vec_f32));
} }
IR::U16 Translator::Convert_F32_to_U16_Normalized(const IR::F32& src) { IR::U32 Translator::Convert_F32_to_U16_Normalized(const IR::F32& src) {
const IR::F32 as_float = ir.FPMul(src, ir.Imm32((f32)std::numeric_limits<u16>::max())); const IR::F32 clamped = ir.FPClamp(src, ir.Imm32(0.0f), ir.Imm32(1.0f));
const IR::F32 as_float = ir.FPMul(clamped, ir.Imm32((f32)std::numeric_limits<u16>::max()));
const IR::U32 as_unsigned = ir.ConvertFToU(32, as_float); const IR::U32 as_unsigned = ir.ConvertFToU(32, as_float);
return ir.UConvert(16, as_unsigned); return as_unsigned;
} }
void Translator::V_CVT_PKNORM_U16_F32(const GcnInst& inst) { void Translator::V_CVT_PKNORM_U16_F32(const GcnInst& inst) {
const IR::VectorReg dst_reg{inst.dst[0].code}; const IR::VectorReg dst_reg{inst.dst[0].code};
const IR::F32 src0 = GetSrc<IR::F32>(inst.src[0]); const IR::U32 src0 = Convert_F32_to_U16_Normalized(GetSrc<IR::F32>(inst.src[0]));
const IR::F32 src1 = GetSrc<IR::F32>(inst.src[1]); const IR::U32 src1 = Convert_F32_to_U16_Normalized(GetSrc<IR::F32>(inst.src[1]));
const IR::Value vec_u16 = ir.SetVectorReg(
ir.CompositeConstruct( dst_reg,
Convert_F32_to_U16_Normalized(src0), ir.BitwiseOr(
Convert_F32_to_U16_Normalized(src1) src0,
); ir.ShiftLeftLogical(src1, ir.Imm32(16))
)
ir.SetVectorReg(dst_reg, ir.PackHalf2x16(vec_u16)); );
} }
void Translator::V_CVT_F32_F16(const GcnInst& inst) { void Translator::V_CVT_F32_F16(const GcnInst& inst) {