first attempt to write inline ASM in C#, failed due to proteted ram

Math/Asm/Asm.sln

@@ -3,7 +3,7 @@ Microsoft Visual Studio Solution File, Format Version 12.00
 # Visual Studio Version 17
 VisualStudioVersion = 17.3.32825.248
 MinimumVisualStudioVersion = 10.0.40219.1
-Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "Asm", "Asm.vcxproj", "{396B053C-CE3C-40A1-AC57-F25F0EEDE910}"
+Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "Asm", "Asm.vcxproj", "{F75765EF-DFF3-4A7A-A7CB-F05993AEE84F}"
 EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
@@ -13,18 +13,19 @@ Global
 		Release|x86 = Release|x86
 	EndGlobalSection
 	GlobalSection(ProjectConfigurationPlatforms) = postSolution
-		{396B053C-CE3C-40A1-AC57-F25F0EEDE910}.Debug|x64.ActiveCfg = Debug|x64
-		{396B053C-CE3C-40A1-AC57-F25F0EEDE910}.Debug|x86.ActiveCfg = Debug|Win32
-		{396B053C-CE3C-40A1-AC57-F25F0EEDE910}.Debug|x86.Build.0 = Debug|Win32
-		{396B053C-CE3C-40A1-AC57-F25F0EEDE910}.Release|x64.ActiveCfg = Release|x64
-		{396B053C-CE3C-40A1-AC57-F25F0EEDE910}.Release|x64.Build.0 = Release|x64
-		{396B053C-CE3C-40A1-AC57-F25F0EEDE910}.Release|x86.ActiveCfg = Release|Win32
-		{396B053C-CE3C-40A1-AC57-F25F0EEDE910}.Release|x86.Build.0 = Release|Win32
+		{F75765EF-DFF3-4A7A-A7CB-F05993AEE84F}.Debug|x64.ActiveCfg = Debug|x64
+		{F75765EF-DFF3-4A7A-A7CB-F05993AEE84F}.Debug|x64.Build.0 = Debug|x64
+		{F75765EF-DFF3-4A7A-A7CB-F05993AEE84F}.Debug|x86.ActiveCfg = Debug|Win32
+		{F75765EF-DFF3-4A7A-A7CB-F05993AEE84F}.Debug|x86.Build.0 = Debug|Win32
+		{F75765EF-DFF3-4A7A-A7CB-F05993AEE84F}.Release|x64.ActiveCfg = Release|x64
+		{F75765EF-DFF3-4A7A-A7CB-F05993AEE84F}.Release|x64.Build.0 = Release|x64
+		{F75765EF-DFF3-4A7A-A7CB-F05993AEE84F}.Release|x86.ActiveCfg = Release|Win32
+		{F75765EF-DFF3-4A7A-A7CB-F05993AEE84F}.Release|x86.Build.0 = Release|Win32
 	EndGlobalSection
 	GlobalSection(SolutionProperties) = preSolution
 		HideSolutionNode = FALSE
 	EndGlobalSection
 	GlobalSection(ExtensibilityGlobals) = postSolution
-		SolutionGuid = {74D9C0A1-8196-4FFA-BCF9-EBC39FA657DB}
+		SolutionGuid = {27162E4C-F467-4555-865D-560856AEF3CD}
 	EndGlobalSection
 EndGlobal

Math/Asm/Asm.vcxproj

@@ -21,7 +21,7 @@
   <PropertyGroup Label="Globals">
     <VCProjectVersion>16.0</VCProjectVersion>
     <Keyword>Win32Proj</Keyword>
-    <ProjectGuid>{396b053c-ce3c-40a1-ac57-f25f0eede910}</ProjectGuid>
+    <ProjectGuid>{f75765ef-dff3-4a7a-a7cb-f05993aee84f}</ProjectGuid>
     <RootNamespace>Asm</RootNamespace>
     <WindowsTargetPlatformVersion>10.0</WindowsTargetPlatformVersion>
   </PropertyGroup>
@@ -81,6 +81,7 @@
     <Link>
       <SubSystem>Console</SubSystem>
       <GenerateDebugInformation>true</GenerateDebugInformation>
+      <EntryPointSymbol>main</EntryPointSymbol>
     </Link>
   </ItemDefinitionGroup>
   <ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
@@ -129,7 +130,7 @@
     </Link>
   </ItemDefinitionGroup>
   <ItemGroup>
-    <MASM Include="源.asm" />
+    <MASM Include="test.asm" />
   </ItemGroup>
   <Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
   <ImportGroup Label="ExtensionTargets">

Math/Asm/Asm.vcxproj.filters

@@ -15,7 +15,7 @@
     </Filter>
   </ItemGroup>
   <ItemGroup>
-    <MASM Include="源.asm">
+    <MASM Include="test.asm">
       <Filter>源文件</Filter>
     </MASM>
   </ItemGroup>

Math/Asm/Debug/Asm.tlog/Asm.lastbuildstate

@@ -0,0 +1,2 @@
+PlatformToolSet=v143:VCToolArchitecture=Native32Bit:VCToolsVersion=14.33.31629:TargetPlatformVersion=10.0.19041.0:
+Debug|Win32|D:\NitloadIndustry\研究项目\1-Flagship\TH2020\ProjectFile\Math\Asm\|

Math/Asm/test.asm

@@ -0,0 +1,21 @@
+.code ; 代码段
+main proc frame
+mov rax, 1235321532151111
+mov rax, 0000000000000000
+mov rax, 1111451111451111
+sub rsp, 20h ; 分配32字节的栈空间，保证16字节对齐
+mov rax,rsp
+mov dword ptr [rax], 25 ; 将25存储在栈顶，作为p的低32位
+movdqu xmm0, [rax] ; 将栈顶的16字节加载到xmm0，作为p的值
+movdqu xmm1, [rax]
+movdqu xmm1, [rsp]
+sqrtss xmm0, xmm0 ; 计算xmm0的低32位的平方根，并存储在xmm1的低32位
+movdqu [rax], xmm0
+movdqu [rax], xmm1
+movdqu [rsp+16], xmm1 ; 将xmm1的值存储在栈顶+16字节的位置，作为q的值
+; 此时，q的值就在[rsp+16]处，可以用其他指令进行操作或返回
+add rsp, 32 ; 恢复栈指针
+.endprolog
+ret
+main endp
+end

Math/Asm/源.asm

@@ -1,9 +0,0 @@
-.code;代码段
-main proc;a函数开始
- mov rax,120
- add rax,34
- sub rax,130
- ;movhpd xmm0,12
- ret
-main endp;a函数结束
-end

源代码/StgSharp.sln

@@ -7,11 +7,6 @@ Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "StgSharp", "StgSharp\StgSha
 EndProject
 Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "StgSharpDebug", "StgSharpDebug\StgSharpDebug.csproj", "{467F2A1A-D7E2-4904-AAAD-A5998EE45C71}"
 EndProject
-Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Dependency", "Dependency", "{063F21D6-CDE3-4DE1-9008-2A211F59AB38}"
-	ProjectSection(SolutionItems) = preProject
-		Dependency\glfw3.dll = Dependency\glfw3.dll
-	EndProjectSection
-EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Any CPU = Debug|Any CPU

源代码/StgSharp/Geometries/Polygon6.cs

@@ -0,0 +1,102 @@
+using Steamworks;
+using StgSharp.Control;
+using StgSharp.Math;
+using System;
+using System.Collections.Generic;
+using System.Numerics;
+using System.Text;
+
+namespace StgSharp.Geometries
+{
+    public class Polygon6 : IPlainGeometry
+    {
+        internal Point _refOrigin;
+        internal Point vertex1;
+        internal Point vertex2;
+        internal Point vertex3;
+        internal Point vertex4;
+        internal Point vertex5;
+        internal Point vertex6;
+        internal readonly int _vertexCount;
+
+        public override Point RefPoint01 => vertex1;
+
+        public override Point RefPoint02 => vertex2;
+
+        public override Point RefPoint03 => vertex3;
+
+
+        internal GetLocationHandler movP01Operation = default;
+        internal GetLocationHandler movP02Operation = default;
+        internal GetLocationHandler movP03Operation = default;
+        internal GetLocationHandler movP04Operation = default;
+        internal GetLocationHandler movP05Operation = default;
+        internal GetLocationHandler movP06Operation = default;
+
+        public GetLocationHandler MovP01Operation { get => movP01Operation; }
+        public GetLocationHandler MovP02Operation { get => movP02Operation; }
+        public GetLocationHandler MovP03Operation { get => movP03Operation; }
+        public GetLocationHandler MovP04Operation { get => movP04Operation; }
+        public GetLocationHandler MovP05Operation { get => movP05Operation; }
+        public GetLocationHandler MovP06Operation { get => movP06Operation; }
+
+        public virtual vec3d movPoint01(uint tick)
+        {
+            return movP01Operation.Invoke(GameTimeLine.tickCounter._value);
+        }
+
+        public virtual vec3d movPoint02(uint tick)
+        {
+            return movP02Operation.Invoke(GameTimeLine.tickCounter._value);
+        }
+
+        public virtual vec3d movPoint03(uint tick)
+        {
+            return movP03Operation.Invoke(GameTimeLine.tickCounter._value);
+        }
+
+        public virtual vec3d movPoint04(uint tick)
+        {
+            return movP04Operation.Invoke(GameTimeLine.tickCounter._value);
+        }
+
+        public virtual vec3d movPoint05(uint tick)
+        {
+            return movP05Operation.Invoke(GameTimeLine.tickCounter._value);
+        }
+
+        public virtual vec3d movPoint06(uint tick)
+        {
+            return movP06Operation.Invoke(GameTimeLine.tickCounter._value);
+        }
+
+        internal override void OnRender(uint tick)
+        {
+            uint nowTick = tick - BornTime;
+            vertex1._position = RefOrigin._position + movPoint01(nowTick);
+            vertex2._position = RefOrigin._position + movPoint02(nowTick);
+            vertex3._position = RefOrigin._position + movPoint03(nowTick);
+            vertex4._position = RefOrigin._position + movPoint04(nowTick);
+            vertex5._position = RefOrigin._position + movPoint05(nowTick);
+            vertex6._position = RefOrigin._position + movPoint06(nowTick);
+        }
+
+        public override Line[] GetAllSides()
+        {
+            return new Line[6]
+                {
+                    new Line(vertex1,vertex2),
+                    new Line(vertex2,vertex3),
+                    new Line(vertex3,vertex4),
+                    new Line(vertex4,vertex5),
+                    new Line(vertex5,vertex6),
+                    new Line(vertex6,vertex1)
+                };
+        }
+
+        public override Plain GetPlain()
+        {
+            return new Plain(vertex1,vertex2,vertex3);
+        }
+    }
+}

源代码/StgSharp/Optimization/CpuInfo.cs

@@ -5,9 +5,9 @@ using System.Text;
 using System.Reflection;
 using System.Runtime.InteropServices;
 
-namespace ssd
+namespace StgSharp.Optimization
 {
-    public class CpuID : IDisposable
+    public class CpuInfo : IDisposable
     {
         [UnmanagedFunctionPointerAttribute(CallingConvention.Cdecl)]
         public delegate void CpuIDDelegate(int level, byte[] buffer);
@@ -68,6 +68,7 @@ namespace ssd
         //    buffer[8] = ecx
         //    buffer[12] = edx
         // }
+        /*
         private byte[] x86CodeBytes =
         {
             0x55,                   // push        ebp  
@@ -90,8 +91,10 @@ namespace ssd
             0x5D,                   // pop         ebp 
             0xc3                    // ret
         };
+        */
 
-        private byte[] x64CodeBytes =
+        //x64
+        private byte[] CodeBytes =
         {
             0x53,                       // push rbx    this gets clobbered by cpuid
 
@@ -117,27 +120,23 @@ namespace ssd
             0xc3                        // ret
         };
 
-        public CpuID()
+        public CpuInfo()
         {
             Compile();
         }
 
-        ~CpuID()
+        ~CpuInfo()
         {
             Dispose(false);
         }
 
         private void Compile()
         {
-            byte[] codeBytes;
+            byte[] codeBytes = CodeBytes;
 
-            if (IntPtr.Size == 4)
+            if (IntPtr.Size != 8)
             {
-                codeBytes = x86CodeBytes;
-            }
-            else
-            {
-                codeBytes = x64CodeBytes;
+                throw new Exception("This class supports 64bits OS only presently");
             }
 
             this.codePointer = VirtualAlloc(

源代码/StgSharp/Optimization/SSE.Convert.cs

@@ -0,0 +1,22 @@
+using StgSharp.Math;
+using System;
+using System.Collections.Generic;
+using System.Text;
+
+namespace StgSharp.Optimization
+{
+    public static unsafe partial class SSE
+    {
+
+        public static m128 ToM128(this vec2d vec)
+        {
+            return *(m128*)&vec;
+        }
+
+        public static m128 ToVec4D(this m128 m)
+        {
+            return *(m128*)&m;
+        }
+
+    }
+}

源代码/StgSharp/Optimization/SSE.SqrtSS.cs

@@ -0,0 +1,83 @@
+using StgSharp.Math;
+using System;
+using System.Collections.Generic;
+using System.Runtime.InteropServices;
+using System.Text;
+using static StgSharp.Optimization.CpuInfo;
+
+namespace StgSharp.Optimization
+{
+    public static unsafe partial class SSE
+    {
+        public static m128 sqrtssM128Cache = new m128();
+
+        private static IntPtr sqrtssCodePointer;
+
+        private static sqrtssDelegate sqrtssDele;
+
+        private static byte[] sqrtssCodeBytes
+            = new byte[]
+        {
+            0x48,0xb8,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,  // mov rax, [m128 location]
+            0xf3,0x0f,0x6f,0x00,                                // movqdu xmm0,xmmword ptr [rax]
+            0xf3,0x0f,0x51,0xc0,                                // sqrtss xmm0,xmm0
+            0xf3,0x0f,0x7f,0x00                                 // movdqu xmmword ptr [rax],xmm0
+        };
+
+        [DllImport("kernel32.dll", SetLastError = true)]
+        private static extern IntPtr VirtualAlloc(
+            IntPtr lpAddress,
+            UIntPtr dwSize,
+            AllocationType flAllocationType,
+            MemoryProtection flProtect
+        );
+
+        [DllImport("kernel32")]
+        private static extern bool VirtualFree(
+                IntPtr lpAddress,
+                UInt32 dwSize,
+                UInt32 dwFreeType
+        );
+
+        private unsafe static void InitSqrtss()
+        {
+            var marshalPtr = Marshal.AllocHGlobal(32);
+            var m128ptr = new IntPtr(16*((long)marshalPtr/16+1));
+            m128* sqrtssCachePointer = (m128*)m128ptr.ToPointer();
+
+            sqrtssM128Cache = *sqrtssCachePointer;
+            byte[] sqrtssM128PointerBytes = BitConverter.GetBytes((ulong)sqrtssCachePointer);
+            for (int i = 0; i < sqrtssM128PointerBytes.Length; i++)
+            {
+                sqrtssCodeBytes[2+i] = sqrtssCodeBytes[7-i];
+            }
+
+            if (IntPtr.Size!=8)
+            {
+                throw new Exception("This SSE class support 64bits OS only presently!");
+            }
+
+            sqrtssCodePointer = VirtualAlloc(
+                IntPtr.Zero,
+                new UIntPtr((uint)sqrtssCodeBytes.Length),
+                AllocationType.COMMIT | AllocationType.RESERVE,
+                MemoryProtection.EXECUTE_READWRITE
+            );
+
+            Marshal.Copy(sqrtssCodeBytes, 0, sqrtssCodePointer, sqrtssCodeBytes.Length);
+            sqrtssDele = (sqrtssDelegate)Marshal.GetDelegateForFunctionPointer(sqrtssCodePointer,typeof(sqrtssDelegate));
+        }
+
+        public static float Sqrtss(float a)
+        {
+            sqrtssM128Cache.float0 = a;
+            sqrtssDele();
+            return sqrtssM128Cache.float0;
+        }
+
+    }
+
+
+    [UnmanagedFunctionPointerAttribute(CallingConvention.StdCall)]
+    internal delegate void sqrtssDelegate();
+}

源代码/StgSharp/Optimization/SSE.cs

@@ -4,9 +4,13 @@ using System.Text;
 
 namespace StgSharp.Optimization
 {
-    public unsafe class SSE
+    public static unsafe partial class SSE
     {
-
+        static SSE()
+        {
+            //编译模拟Sqrtss指令
+            InitSqrtss();
+        }
 
     }
 }

源代码/StgSharp/obj/Debug/netstandard2.1/StgSharp.csproj.CoreCompileInputs.cache

@@ -1 +1 @@
-0bd409a23d12a7e03a996538d35e30898336817d
+508c7d643da3ffba0187af6df00e8fd1f0a8c855

源代码/StgSharpDebug/Program.cs

@@ -40,11 +40,8 @@ internal class Program
         }
         */
 
-    }
+        Console.WriteLine(StgSharp.Optimization.SSE.Sqrtss(25.0f));
 
-    public class MyNum : INode<MyNum>
-    {
-        
     }
 }
     

源代码/StgSharpDebug/bin/Debug/net6.0/StgSharp.xml

@@ -4,35 +4,65 @@
         <name>StgSharp</name>
     </assembly>
     <members>
-        <member name="P:StgSharp.Entity.CollisionSensor">
+        <member name="P:StgSharp.IEntity.CollisionSensor">
             <summary>
             A sensor used for cheking out if a bullet hit a player or something else
             </summary>
         </member>
-        <member name="P:StgSharp.Entity.Location">
+        <member name="P:StgSharp.IEntity.Location">
             <summary>
             A 2D vector presenting the position of an enemy entity, position can also be visited by X, Y
             </summary>
         </member>
-        <member name="P:StgSharp.Entity.X">
+        <member name="P:StgSharp.IEntity.X">
             <summary>
             X position of an enenmy entity
             </summary>
         </member>
-        <member name="P:StgSharp.Entity.Y">
+        <member name="P:StgSharp.IEntity.Y">
             <summary>
             Y position of an enemy entity
             </summary>
         </member>
-        <member name="F:StgSharp.Geometries.RenderMode.BasePoint">
+        <member name="T:StgSharp.Geometries.UnusedVertexException">
             <summary>
-            坐标点将基于参考原点构建
+            当几何体调用了过多的CalcVec方法时，产生UnusedVertexException异常
             </summary>
         </member>
-        <member name="F:StgSharp.Geometries.RenderMode.RefPoint">
+        <member name="T:StgSharp.Geometries.GetLocationHandler">
             <summary>
-            坐标点将基于参考起点构建
+            计算位移失量，需要手动输入游戏刻以提供时间数据
             </summary>
+            <param name="tickCounter"></param>
+            <returns>表示位移量的三维矢量</returns>
+        </member>
+        <member name="T:StgSharp.Geometries.IPlainGeometry">
+            <summary>
+            平面几何体。包含三角形，对称四边形，自由多边形等。最多可由16个点参数定义。
+            </summary>
+        </member>
+        <member name="M:StgSharp.Geometries.Parallelogram.MovVertex03(System.UInt32)">
+            <summary>
+            移动平行四边形的第三个顶点，是一个多余的方法，调用会引发多余变量异常
+            </summary>
+            <param name="tick">当前游戏刻</param>
+            <returns></returns>
+            <exception cref="T:StgSharp.Geometries.UnusedVertexException">几何体过定义</exception>
+        </member>
+        <member name="M:StgSharp.Geometries.Point.Update(System.UInt32)">
+            <summary>
+            刷新点的位置，即移动点。当该点是某个几何体的从动点时
+            （即点坐标由几何体其他点约束或点参数已经在创建几何体时定义）
+            请不要再次进行定义，程序将自行忽略此处的代码。
+            </summary>
+            <param name="tick">当前游戏刻</param>
+        </member>
+        <member name="M:StgSharp.Geometries.Rectangle.MovCenter(System.UInt32)">
+            <summary>
+            计算中心点相对于参考原点的位移
+            </summary>
+            <param name="tick">当前游戏刻</param>
+            <returns></returns>
         </member>
         <member name="M:StgSharp.Math.VectorCalc.multiple(StgSharp.Math.colum2,StgSharp.Math.vec2d)">
             <summary>

源代码/StgSharpDebug/obj/Debug/net6.0/StgSharpDebug.csproj.CoreCompileInputs.cache

@@ -1 +1 @@
-69eea3b78f7825fc761f43509d32b318b370c0d6
+88f0e007daf41ad68341e491760b7799a6c6cbcc