aboutsummaryrefslogtreecommitdiff
path: root/Ryujinx.Graphics/Shader/Translation/Translator.cs
diff options
context:
space:
mode:
authorgdkchan <gab.dark.100@gmail.com>2019-04-17 20:57:08 -0300
committerjduncanator <1518948+jduncanator@users.noreply.github.com>2019-04-18 09:57:08 +1000
commit6b23a2c125b9c48b5ebea92716004ef68698bb0f (patch)
tree69332df6fbbd8e2bddc522ba682fcc5c7a69e101 /Ryujinx.Graphics/Shader/Translation/Translator.cs
parentb2e88b04a85b41cc60af3485d88c90429e84a218 (diff)
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
Diffstat (limited to 'Ryujinx.Graphics/Shader/Translation/Translator.cs')
-rw-r--r--Ryujinx.Graphics/Shader/Translation/Translator.cs219
1 files changed, 219 insertions, 0 deletions
diff --git a/Ryujinx.Graphics/Shader/Translation/Translator.cs b/Ryujinx.Graphics/Shader/Translation/Translator.cs
new file mode 100644
index 00000000..706f3cfa
--- /dev/null
+++ b/Ryujinx.Graphics/Shader/Translation/Translator.cs
@@ -0,0 +1,219 @@
+using Ryujinx.Graphics.Gal;
+using Ryujinx.Graphics.Shader.CodeGen.Glsl;
+using Ryujinx.Graphics.Shader.Decoders;
+using Ryujinx.Graphics.Shader.Instructions;
+using Ryujinx.Graphics.Shader.IntermediateRepresentation;
+using Ryujinx.Graphics.Shader.StructuredIr;
+using Ryujinx.Graphics.Shader.Translation.Optimizations;
+using System.Collections.Generic;
+
+using static Ryujinx.Graphics.Shader.IntermediateRepresentation.OperandHelper;
+
+namespace Ryujinx.Graphics.Shader.Translation
+{
+ public static class Translator
+ {
+ public static ShaderProgram Translate(IGalMemory memory, ulong address, ShaderConfig config)
+ {
+ return Translate(memory, address, 0, config);
+ }
+
+ public static ShaderProgram Translate(
+ IGalMemory memory,
+ ulong address,
+ ulong addressB,
+ ShaderConfig config)
+ {
+ Operation[] shaderOps = DecodeShader(memory, address, config.Type);
+
+ if (addressB != 0)
+ {
+ //Dual vertex shader.
+ Operation[] shaderOpsB = DecodeShader(memory, addressB, config.Type);
+
+ shaderOps = Combine(shaderOps, shaderOpsB);
+ }
+
+ BasicBlock[] irBlocks = ControlFlowGraph.MakeCfg(shaderOps);
+
+ Dominance.FindDominators(irBlocks[0], irBlocks.Length);
+
+ Dominance.FindDominanceFrontiers(irBlocks);
+
+ Ssa.Rename(irBlocks);
+
+ Optimizer.Optimize(irBlocks);
+
+ StructuredProgramInfo sInfo = StructuredProgram.MakeStructuredProgram(irBlocks);
+
+ GlslProgram program = GlslGenerator.Generate(sInfo, config);
+
+ ShaderProgramInfo spInfo = new ShaderProgramInfo(
+ program.CBufferDescriptors,
+ program.TextureDescriptors);
+
+ return new ShaderProgram(spInfo, program.Code);
+ }
+
+ private static Operation[] DecodeShader(IGalMemory memory, ulong address, GalShaderType shaderType)
+ {
+ ShaderHeader header = new ShaderHeader(memory, address);
+
+ Block[] cfg = Decoder.Decode(memory, address);
+
+ EmitterContext context = new EmitterContext(shaderType, header);
+
+ for (int blkIndex = 0; blkIndex < cfg.Length; blkIndex++)
+ {
+ Block block = cfg[blkIndex];
+
+ context.CurrBlock = block;
+
+ context.MarkLabel(context.GetLabel(block.Address));
+
+ for (int opIndex = 0; opIndex < block.OpCodes.Count; opIndex++)
+ {
+ OpCode op = block.OpCodes[opIndex];
+
+ if (op.NeverExecute)
+ {
+ continue;
+ }
+
+ Operand predSkipLbl = null;
+
+ bool skipPredicateCheck = op.Emitter == InstEmit.Bra;
+
+ if (op is OpCodeSync opSync)
+ {
+ //If the instruction is a SYNC instruction with only one
+ //possible target address, then the instruction is basically
+ //just a simple branch, we can generate code similar to branch
+ //instructions, with the condition check on the branch itself.
+ skipPredicateCheck |= opSync.Targets.Count < 2;
+ }
+
+ if (!(op.Predicate.IsPT || skipPredicateCheck))
+ {
+ Operand label;
+
+ if (opIndex == block.OpCodes.Count - 1 && block.Next != null)
+ {
+ label = context.GetLabel(block.Next.Address);
+ }
+ else
+ {
+ label = Label();
+
+ predSkipLbl = label;
+ }
+
+ Operand pred = Register(op.Predicate);
+
+ if (op.InvertPredicate)
+ {
+ context.BranchIfTrue(label, pred);
+ }
+ else
+ {
+ context.BranchIfFalse(label, pred);
+ }
+ }
+
+ context.CurrOp = op;
+
+ op.Emitter(context);
+
+ if (predSkipLbl != null)
+ {
+ context.MarkLabel(predSkipLbl);
+ }
+ }
+ }
+
+ return context.GetOperations();
+ }
+
+ private static Operation[] Combine(Operation[] a, Operation[] b)
+ {
+ //Here we combine two shaders.
+ //For shader A:
+ //- All user attribute stores on shader A are turned into copies to a
+ //temporary variable. It's assumed that shader B will consume them.
+ //- All return instructions are turned into branch instructions, the
+ //branch target being the start of the shader B code.
+ //For shader B:
+ //- All user attribute loads on shader B are turned into copies from a
+ //temporary variable, as long that attribute is written by shader A.
+ List<Operation> output = new List<Operation>(a.Length + b.Length);
+
+ Operand[] temps = new Operand[AttributeConsts.UserAttributesCount * 4];
+
+ Operand lblB = Label();
+
+ for (int index = 0; index < a.Length; index++)
+ {
+ Operation operation = a[index];
+
+ if (IsUserAttribute(operation.Dest))
+ {
+ int tIndex = (operation.Dest.Value - AttributeConsts.UserAttributeBase) / 4;
+
+ Operand temp = temps[tIndex];
+
+ if (temp == null)
+ {
+ temp = Local();
+
+ temps[tIndex] = temp;
+ }
+
+ operation.Dest = temp;
+ }
+
+ if (operation.Inst == Instruction.Return)
+ {
+ output.Add(new Operation(Instruction.Branch, lblB));
+ }
+ else
+ {
+ output.Add(operation);
+ }
+ }
+
+ output.Add(new Operation(Instruction.MarkLabel, lblB));
+
+ for (int index = 0; index < b.Length; index++)
+ {
+ Operation operation = b[index];
+
+ for (int srcIndex = 0; srcIndex < operation.SourcesCount; srcIndex++)
+ {
+ Operand src = operation.GetSource(srcIndex);
+
+ if (IsUserAttribute(src))
+ {
+ Operand temp = temps[(src.Value - AttributeConsts.UserAttributeBase) / 4];
+
+ if (temp != null)
+ {
+ operation.SetSource(srcIndex, temp);
+ }
+ }
+ }
+
+ output.Add(operation);
+ }
+
+ return output.ToArray();
+ }
+
+ private static bool IsUserAttribute(Operand operand)
+ {
+ return operand != null &&
+ operand.Type == OperandType.Attribute &&
+ operand.Value >= AttributeConsts.UserAttributeBase &&
+ operand.Value < AttributeConsts.UserAttributeEnd;
+ }
+ }
+} \ No newline at end of file
generated by cgit v1.2.3-70-g09d2 (git 2.48.1) at 2025-03-13 18:12:57 +0000