vleeuwenmenno/Ryujinx
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add Profiled Persistent Translation Cache. (#769)

Browse Source 
* Delete DelegateTypes.cs

* Delete DelegateCache.cs

* Add files via upload

* Update Horizon.cs

* Update Program.cs

* Update MainWindow.cs

* Update Aot.cs

* Update RelocEntry.cs

* Update Translator.cs

* Update MemoryManager.cs

* Update InstEmitMemoryHelper.cs

* Update Delegates.cs

* Nit.

* Nit.

* Nit.

* 10 fewer MSIL bytes for us

* Add comment. Nits.

* Update Translator.cs

* Update Aot.cs

* Nits.

* Opt..

* Opt..

* Opt..

* Opt..

* Allow to change compression level.

* Update MemoryManager.cs

* Update Translator.cs

* Manage corner cases during the save phase. Nits.

* Update Aot.cs

* Translator response tweak for Aot disabled. Nit.

* Nit.

* Nits.

* Create DelegateHelpers.cs

* Update Delegates.cs

* Nit.

* Nit.

* Nits.

* Fix due to #784.

* Fixes due to #757 & #841.

* Fix due to #846.

* Fix due to #847.

* Use MethodInfo for managed method calls.

Use IR methods instead of managed methods about Max/Min (S/U).
Follow-ups & Nits.

* Add missing exception messages.

Reintroduce slow path for Fmov_Vi.
Implement slow path for Fmov_Si.

* Switch to the new folder structure.

Nits.

* Impl. index-based relocation information. Impl. cache file version field.

* Nit.

* Address gdkchan comments.

Mainly:
- fixed cache file corruption issue on exit; - exposed a way to disable AOT on the GUI.

* Address AcK77 comment.

* Address Thealexbarney, jduncanator & emmauss comments.

Header magic, CpuId (FI) & Aot -> Ptc.

* Adaptation to the new application reloading system.

Improvements to the call system of managed methods.
Follow-ups.
Nits.

* Get the same boot times as on master when PTC is disabled.

* Profiled Aot.

* A32 support (#897).

* #975 support (1 of 2).

* #975 support (2 of 2).

* Rebase fix & nits.

* Some fixes and nits (still one bug left).

* One fix & nits.

* Tests fix (by gdk) & nits.

* Support translations not only in high quality and rejit.

Nits.

* Added possibility to skip translations and continue execution, using `ESC` key.

* Update SettingsWindow.cs

* Update GLRenderer.cs

* Update Ptc.cs

* Disabled Profiled PTC by default as requested in the past by gdk.

* Fix rejit bug. Increased number of parallel translations. Add stack unwinding stuffs support (1 of 2).

Nits.

* Add stack unwinding stuffs support (2 of 2). Tuned number of parallel translations.

* Restored the ability to assemble jumps with 8-bit offset when Profiled PTC is disabled or during profiling.

Modifications due to rebase.
Nits.

* Limited profiling of the functions to be translated to the addresses belonging to the range of static objects only.

* Nits.

* Nits.

* Update Delegates.cs

* Nit.

* Update InstEmitSimdArithmetic.cs

* Address riperiperi comments.

* Fixed the issue of unjustifiably longer boot times at the second boot than at the first boot, measured at the same time or reference point and with the same number of translated functions.

* Implemented a simple redundant load/save mechanism.

Halved the value of Decoder.MaxInstsPerFunction more appropriate for the current performance of the Translator.
Replaced by Logger.PrintError to Logger.PrintDebug in TexturePool.cs about the supposed invalid texture format to avoid the spawn of the log.
Nits.

* Nit.

Improved Logger.PrintError in TexturePool.cs to avoid log spawn.
Added missing code for FZ handling (in output) for fp max/min instructions (slow paths).

* Add configuration migration for PTC

Co-authored-by: Thog <me@thog.eu>
This commit is contained in:
LDj3SNuD
2020-06-16 20:28:02 +02:00
committed by GitHub
parent fa286d3535
commit 5e724cf24e
78 changed files with 3066 additions and 1207 deletions
Download Patch File Download Diff File Expand all files Collapse all files

98
ARMeilleure/Instructions/InstEmitSimdCvt32.cs
View File

@ -4,6 +4,7 @@ using ARMeilleure.State;
using ARMeilleure.Translation;
using System;
using System.Diagnostics;
using System.Reflection;
using static ARMeilleure.Instructions.InstEmitHelper;
using static ARMeilleure.Instructions.InstEmitSimdHelper;
@ -20,7 +21,7 @@ namespace ARMeilleure.Instructions
{
// Move the low bit to the top.
return ((vd & 0x1) << 4) | (vd >> 1);
}
}
else
{
// Move the high bit to the bottom.
@ -30,29 +31,22 @@ namespace ARMeilleure.Instructions
private static Operand EmitSaturateFloatToInt(ArmEmitterContext context, Operand op1, bool unsigned)
{
MethodInfo info;
if (op1.Type == OperandType.FP64)
{
if (unsigned)
{
return context.Call(new _U32_F64(SoftFallback.SatF64ToU32), op1);
}
else
{
return context.Call(new _S32_F64(SoftFallback.SatF64ToS32), op1);
}
info = unsigned
? typeof(SoftFallback).GetMethod(nameof(SoftFallback.SatF64ToU32))
: typeof(SoftFallback).GetMethod(nameof(SoftFallback.SatF64ToS32));
}
else
{
if (unsigned)
{
return context.Call(new _U32_F32(SoftFallback.SatF32ToU32), op1);
}
else
{
return context.Call(new _S32_F32(SoftFallback.SatF32ToS32), op1);
}
info = unsigned
? typeof(SoftFallback).GetMethod(nameof(SoftFallback.SatF32ToU32))
: typeof(SoftFallback).GetMethod(nameof(SoftFallback.SatF32ToS32));
}
return context.Call(info, op1);
}
public static void Vcvt_V(ArmEmitterContext context)
@ -96,7 +90,7 @@ namespace ARMeilleure.Instructions
res2 = context.AddIntrinsic(Intrinsic.X86Cvtdq2ps, res2);
return context.AddIntrinsic(Intrinsic.X86Addps, res, res2);
}
}
else
{
return context.AddIntrinsic(Intrinsic.X86Cvtdq2ps, n);
@ -114,9 +108,7 @@ namespace ARMeilleure.Instructions
EmitVectorUnaryOpSx32(context, (op1) => EmitFPConvert(context, op1, floatSize, true));
}
}
}
}
public static void Vcvt_FD(ArmEmitterContext context)
@ -173,29 +165,22 @@ namespace ARMeilleure.Instructions
// TODO: Fast Path.
if (roundWithFpscr)
{
MethodInfo info;
if (floatSize == OperandType.FP64)
{
if (unsigned)
{
asInteger = context.Call(new _U32_F64(SoftFallback.DoubleToUInt32), toConvert);
}
else
{
asInteger = context.Call(new _S32_F64(SoftFallback.DoubleToInt32), toConvert);
}
info = unsigned
? typeof(SoftFallback).GetMethod(nameof(SoftFallback.DoubleToUInt32))
: typeof(SoftFallback).GetMethod(nameof(SoftFallback.DoubleToInt32));
}
else
{
if (unsigned)
{
asInteger = context.Call(new _U32_F32(SoftFallback.FloatToUInt32), toConvert);
}
else
{
asInteger = context.Call(new _S32_F32(SoftFallback.FloatToInt32), toConvert);
}
info = unsigned
? typeof(SoftFallback).GetMethod(nameof(SoftFallback.FloatToUInt32))
: typeof(SoftFallback).GetMethod(nameof(SoftFallback.FloatToInt32));
}
asInteger = context.Call(info, toConvert);
}
else
{
@ -205,7 +190,7 @@ namespace ARMeilleure.Instructions
InsertScalar(context, op.Vd, asInteger);
}
}
}
else
{
bool unsigned = op.Opc == 0;
@ -218,22 +203,17 @@ namespace ARMeilleure.Instructions
}
}
public static Operand EmitRoundMathCall(ArmEmitterContext context, MidpointRounding roundMode, Operand n)
private static Operand EmitRoundMathCall(ArmEmitterContext context, MidpointRounding roundMode, Operand n)
{
IOpCode32Simd op = (IOpCode32Simd)context.CurrOp;
Delegate dlg;
string name = nameof(Math.Round);
if ((op.Size & 1) == 0)
{
dlg = new _F32_F32_MidpointRounding(MathF.Round);
}
else /* if ((op.Size & 1) == 1) */
{
dlg = new _F64_F64_MidpointRounding(Math.Round);
}
MethodInfo info = (op.Size & 1) == 0
? typeof(MathF).GetMethod(name, new Type[] { typeof(float), typeof(MidpointRounding) })
: typeof(Math). GetMethod(name, new Type[] { typeof(double), typeof(MidpointRounding) });
return context.Call(dlg, n, Const((int)roundMode));
return context.Call(info, n, Const((int)roundMode));
}
private static FPRoundingMode RMToRoundMode(int rm)
@ -282,10 +262,10 @@ namespace ARMeilleure.Instructions
toConvert = EmitRoundMathCall(context, MidpointRounding.ToEven, toConvert);
break;
case 0b10: // Towards positive infinity
toConvert = EmitUnaryMathCall(context, MathF.Ceiling, Math.Ceiling, toConvert);
toConvert = EmitUnaryMathCall(context, nameof(Math.Ceiling), toConvert);
break;
case 0b11: // Towards negative infinity
toConvert = EmitUnaryMathCall(context, MathF.Floor, Math.Floor, toConvert);
toConvert = EmitUnaryMathCall(context, nameof(Math.Floor), toConvert);
break;
}
@ -316,7 +296,7 @@ namespace ARMeilleure.Instructions
return context.AddIntrinsic(inst, m, Const(X86GetRoundControl(roundMode)));
});
}
else
else
{
Operand toConvert = ExtractScalar(context, floatSize, op.Vm);
@ -329,10 +309,10 @@ namespace ARMeilleure.Instructions
toConvert = EmitRoundMathCall(context, MidpointRounding.ToEven, toConvert);
break;
case 0b10: // Towards positive infinity
toConvert = EmitUnaryMathCall(context, MathF.Ceiling, Math.Ceiling, toConvert);
toConvert = EmitUnaryMathCall(context, nameof(Math.Ceiling), toConvert);
break;
case 0b11: // Towards negative infinity
toConvert = EmitUnaryMathCall(context, MathF.Floor, Math.Floor, toConvert);
toConvert = EmitUnaryMathCall(context, nameof(Math.Floor), toConvert);
break;
}
@ -351,10 +331,10 @@ namespace ARMeilleure.Instructions
Intrinsic inst = (op.Size & 1) == 0 ? Intrinsic.X86Roundss : Intrinsic.X86Roundsd;
return context.AddIntrinsic(inst, m, Const(X86GetRoundControl(FPRoundingMode.TowardsZero)));
});
}
}
else
{
EmitScalarUnaryOpF32(context, (op1) => EmitUnaryMathCall(context, MathF.Truncate, Math.Truncate, op1));
EmitScalarUnaryOpF32(context, (op1) => EmitUnaryMathCall(context, nameof(Math.Truncate), op1));
}
}
@ -423,7 +403,7 @@ namespace ARMeilleure.Instructions
if (signed)
{
dRes = context.BitwiseExclusiveOr(nIntOrLong, nInt);
}
}
else
{
dRes = context.BitwiseExclusiveOr(nIntOrLong2, nInt);
@ -527,7 +507,7 @@ namespace ARMeilleure.Instructions
if (signed)
{
return context.AddIntrinsic(Intrinsic.X86Pxor, nInt, nRes);
}
}
else
{
Operand dRes = context.AddIntrinsic(Intrinsic.X86Pxor, nInt2, nRes);