More unrolling attempts

Author: Nick-Nuon

.vscode/launch.json

@@ -0,0 +1,26 @@
+{
+    "version": "0.2.0",
+    "configurations": [
+        {
+            // Use IntelliSense to find out which attributes exist for C# debugging
+            // Use hover for the description of the existing attributes
+            // For further information visit https://github.com/dotnet/vscode-csharp/blob/main/debugger-launchjson.md
+            "name": ".NET Core Launch (console)",
+            "type": "coreclr",
+            "request": "launch",
+            "preLaunchTask": "build",
+            // If you have changed target frameworks, make sure to update the program path.
+            "program": "${workspaceFolder}/benchmark/bin/Debug/net7.0/benchmark.dll",
+            "args": [],
+            "cwd": "${workspaceFolder}/benchmark",
+            // For more information about the 'console' field, see https://aka.ms/VSCode-CS-LaunchJson-Console
+            "console": "internalConsole",
+            "stopAtEntry": false
+        },
+        {
+            "name": ".NET Core Attach",
+            "type": "coreclr",
+            "request": "attach"
+        }
+    ]
+}

.vscode/tasks.json

@@ -0,0 +1,41 @@
+{
+    "version": "2.0.0",
+    "tasks": [
+        {
+            "label": "build",
+            "command": "dotnet",
+            "type": "process",
+            "args": [
+                "build",
+                "${workspaceFolder}/benchmark/benchmark.csproj",
+                "/property:GenerateFullPaths=true",
+                "/consoleloggerparameters:NoSummary;ForceNoAlign"
+            ],
+            "problemMatcher": "$msCompile"
+        },
+        {
+            "label": "publish",
+            "command": "dotnet",
+            "type": "process",
+            "args": [
+                "publish",
+                "${workspaceFolder}/benchmark/benchmark.csproj",
+                "/property:GenerateFullPaths=true",
+                "/consoleloggerparameters:NoSummary;ForceNoAlign"
+            ],
+            "problemMatcher": "$msCompile"
+        },
+        {
+            "label": "watch",
+            "command": "dotnet",
+            "type": "process",
+            "args": [
+                "watch",
+                "run",
+                "--project",
+                "${workspaceFolder}/benchmark/benchmark.csproj"
+            ],
+            "problemMatcher": "$msCompile"
+        }
+    ]
+}

src/Ascii.cs

@@ -119,7 +119,7 @@ public static unsafe bool SIMDIsAscii(this ReadOnlySpan<char> s)
                         } */
 
 
-/* 
+/* Unrolled twice:
 |                 Method |   N |       Mean |    Error |   StdDev |
 |----------------------- |---- |-----------:|---------:|---------:|
 |     FastUnicodeIsAscii | 100 |   905.7 ns | 17.95 ns | 20.67 ns |
@@ -144,9 +144,43 @@ public static unsafe bool SIMDIsAscii(this ReadOnlySpan<char> s)
                             Vector128<ushort> raw2 = Sse41.LoadDquVector128((ushort*)pStart + i + 8);
 
                             total = Sse2.Or(total, raw1);
-                            total = Sse2.Or(total, raw2);
+                            total = Sse2.Or(total, raw2); 
                         }
 
+// |                 Method |   N |       Mean |     Error |    StdDev |
+// |----------------------- |---- |-----------:|----------:|----------:|
+// |     FastUnicodeIsAscii | 100 | 1,601.3 ns |  31.62 ns |  31.05 ns |
+// | StandardUnicodeIsAscii | 100 | 2,502.5 ns |  49.20 ns |  65.68 ns |
+// |         RuntimeIsAscii | 100 | 2,478.5 ns |  30.08 ns |  26.66 ns |
+// |     FastUnicodeIsAscii | 200 |   653.0 ns |   6.26 ns |   5.86 ns |
+// | StandardUnicodeIsAscii | 200 | 5,282.7 ns | 102.28 ns | 105.03 ns |
+// |         RuntimeIsAscii | 200 | 5,366.1 ns |  65.50 ns |  61.27 ns |
+// |     FastUnicodeIsAscii | 500 | 1,305.4 ns |  11.85 ns |  11.09 ns |
+// | StandardUnicodeIsAscii | 500 | 6,235.6 ns | 103.06 ns |  96.40 ns |
+// |         RuntimeIsAscii | 500 | 6,389.6 ns | 103.20 ns |  96.53 ns |
+
+
+                    // if (s.Length > 32)  // Adjusted for the 4x unrolled loop
+                    // {
+                    //     Vector128<ushort> total = Sse41.LoadDquVector128((ushort*)pStart);
+                    //     i += 8;
+
+                    //     // 4x loop unrolling
+                    //     for (; i + 31 < s.Length; i += 32)
+                    //     {
+                    //         Vector128<ushort> raw1 = Sse41.LoadDquVector128((ushort*)pStart + i);
+                    //         Vector128<ushort> raw2 = Sse41.LoadDquVector128((ushort*)pStart + i + 8);
+                    //         Vector128<ushort> raw3 = Sse41.LoadDquVector128((ushort*)pStart + i + 16);
+                    //         Vector128<ushort> raw4 = Sse41.LoadDquVector128((ushort*)pStart + i + 24);
+                            
+                    //         total = Sse2.Or(total, raw1);
+                    //         total = Sse2.Or(total, raw2);
+                    //         total = Sse2.Or(total, raw3);
+                    //         total = Sse2.Or(total, raw4);
+                    //     }
+                    
+
+
                         Vector128<ushort> b127 = Vector128.Create((ushort)127);
                         Vector128<ushort> b = Sse41.Max(b127, total);
                         Vector128<ushort> b16 = Sse41.CompareEqual(b, b127);