New shader translator implementation (#654)
* Start implementing a new shader translator
* Fix shift instructions and a typo
* Small refactoring on StructuredProgram, move RemovePhis method to a separate class
* Initial geometry shader support
* Implement TLD4
* Fix -- There's no negation on FMUL32I
* Add constant folding and algebraic simplification optimizations, nits
* Some leftovers from constant folding
* Avoid cast for constant assignments
* Add a branch elimination pass, and misc small fixes
* Remove redundant branches, add expression propagation and other improvements on the code
* Small leftovers -- add missing break and continue, remove unused properties, other improvements
* Add null check to handle empty block cases on block visitor
* Add HADD2 and HMUL2 half float shader instructions
* Optimize pack/unpack sequences, some fixes related to half float instructions
* Add TXQ, TLD, TLDS and TLD4S shader texture instructions, and some support for bindless textures, some refactoring on codegen
* Fix copy paste mistake that caused RZ to be ignored on the AST instruction
* Add workaround for conditional exit, and fix half float instruction with constant buffer
* Add missing 0.0 source for TLDS.LZ variants
* Simplify the switch for TLDS.LZ
* Texture instructions related fixes
* Implement the HFMA instruction, and some misc. fixes
* Enable constant folding on UnpackHalf2x16 instructions
* Refactor HFMA to use OpCode* for opcode decoding rather than on the helper methods
* Remove the old shader translator
* Remove ShaderDeclInfo and other unused things
* Add dual vertex shader support
* Add ShaderConfig, used to pass shader type and maximum cbuffer size
* Move and rename some instruction enums
* Move texture instructions into a separate file
* Move operand GetExpression and locals management to OperandManager
* Optimize opcode decoding using a simple list and binary search
* Add missing condition for do-while on goto elimination
* Misc. fixes on texture instructions
* Simplify TLDS switch
* Address PR feedback, and a nit
2019-04-17 23:57:08 +00:00
|
|
|
using Ryujinx.Graphics.Shader.Decoders;
|
|
|
|
using Ryujinx.Graphics.Shader.IntermediateRepresentation;
|
|
|
|
using Ryujinx.Graphics.Shader.Translation;
|
|
|
|
using System;
|
|
|
|
|
|
|
|
using static Ryujinx.Graphics.Shader.Instructions.InstEmitHelper;
|
|
|
|
using static Ryujinx.Graphics.Shader.IntermediateRepresentation.OperandHelper;
|
|
|
|
|
|
|
|
namespace Ryujinx.Graphics.Shader.Instructions
|
|
|
|
{
|
|
|
|
static class InstEmitAluHelper
|
|
|
|
{
|
|
|
|
public static int GetIntMin(IntegerType type)
|
|
|
|
{
|
|
|
|
switch (type)
|
|
|
|
{
|
|
|
|
case IntegerType.U8: return byte.MinValue;
|
|
|
|
case IntegerType.S8: return sbyte.MinValue;
|
|
|
|
case IntegerType.U16: return ushort.MinValue;
|
|
|
|
case IntegerType.S16: return short.MinValue;
|
|
|
|
case IntegerType.U32: return (int)uint.MinValue;
|
|
|
|
case IntegerType.S32: return int.MinValue;
|
|
|
|
}
|
|
|
|
|
|
|
|
throw new ArgumentException($"The type \"{type}\" is not a supported int type.");
|
|
|
|
}
|
|
|
|
|
|
|
|
public static int GetIntMax(IntegerType type)
|
|
|
|
{
|
|
|
|
switch (type)
|
|
|
|
{
|
|
|
|
case IntegerType.U8: return byte.MaxValue;
|
|
|
|
case IntegerType.S8: return sbyte.MaxValue;
|
|
|
|
case IntegerType.U16: return ushort.MaxValue;
|
|
|
|
case IntegerType.S16: return short.MaxValue;
|
|
|
|
case IntegerType.U32: return unchecked((int)uint.MaxValue);
|
|
|
|
case IntegerType.S32: return int.MaxValue;
|
|
|
|
}
|
|
|
|
|
|
|
|
throw new ArgumentException($"The type \"{type}\" is not a supported int type.");
|
|
|
|
}
|
|
|
|
|
|
|
|
public static Operand GetPredLogicalOp(
|
|
|
|
EmitterContext context,
|
|
|
|
LogicalOperation logicalOp,
|
|
|
|
Operand input,
|
|
|
|
Operand pred)
|
|
|
|
{
|
|
|
|
switch (logicalOp)
|
|
|
|
{
|
|
|
|
case LogicalOperation.And: return context.BitwiseAnd (input, pred);
|
|
|
|
case LogicalOperation.Or: return context.BitwiseOr (input, pred);
|
|
|
|
case LogicalOperation.ExclusiveOr: return context.BitwiseExclusiveOr(input, pred);
|
|
|
|
}
|
|
|
|
|
|
|
|
return input;
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void SetZnFlags(EmitterContext context, Operand dest, bool setCC, bool extended = false)
|
|
|
|
{
|
|
|
|
if (!setCC)
|
|
|
|
{
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (extended)
|
|
|
|
{
|
2019-07-02 02:39:22 +00:00
|
|
|
// When the operation is extended, it means we are doing
|
|
|
|
// the operation on a long word with any number of bits,
|
|
|
|
// so we need to AND the zero flag from result with the
|
|
|
|
// previous result when extended is specified, to ensure
|
|
|
|
// we have ZF set only if all words are zero, and not just
|
|
|
|
// the last one.
|
New shader translator implementation (#654)
* Start implementing a new shader translator
* Fix shift instructions and a typo
* Small refactoring on StructuredProgram, move RemovePhis method to a separate class
* Initial geometry shader support
* Implement TLD4
* Fix -- There's no negation on FMUL32I
* Add constant folding and algebraic simplification optimizations, nits
* Some leftovers from constant folding
* Avoid cast for constant assignments
* Add a branch elimination pass, and misc small fixes
* Remove redundant branches, add expression propagation and other improvements on the code
* Small leftovers -- add missing break and continue, remove unused properties, other improvements
* Add null check to handle empty block cases on block visitor
* Add HADD2 and HMUL2 half float shader instructions
* Optimize pack/unpack sequences, some fixes related to half float instructions
* Add TXQ, TLD, TLDS and TLD4S shader texture instructions, and some support for bindless textures, some refactoring on codegen
* Fix copy paste mistake that caused RZ to be ignored on the AST instruction
* Add workaround for conditional exit, and fix half float instruction with constant buffer
* Add missing 0.0 source for TLDS.LZ variants
* Simplify the switch for TLDS.LZ
* Texture instructions related fixes
* Implement the HFMA instruction, and some misc. fixes
* Enable constant folding on UnpackHalf2x16 instructions
* Refactor HFMA to use OpCode* for opcode decoding rather than on the helper methods
* Remove the old shader translator
* Remove ShaderDeclInfo and other unused things
* Add dual vertex shader support
* Add ShaderConfig, used to pass shader type and maximum cbuffer size
* Move and rename some instruction enums
* Move texture instructions into a separate file
* Move operand GetExpression and locals management to OperandManager
* Optimize opcode decoding using a simple list and binary search
* Add missing condition for do-while on goto elimination
* Misc. fixes on texture instructions
* Simplify TLDS switch
* Address PR feedback, and a nit
2019-04-17 23:57:08 +00:00
|
|
|
Operand oldZF = GetZF(context);
|
|
|
|
|
|
|
|
Operand res = context.BitwiseAnd(context.ICompareEqual(dest, Const(0)), oldZF);
|
|
|
|
|
|
|
|
context.Copy(GetZF(context), res);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
context.Copy(GetZF(context), context.ICompareEqual(dest, Const(0)));
|
|
|
|
}
|
|
|
|
|
|
|
|
context.Copy(GetNF(context), context.ICompareLess(dest, Const(0)));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|