....

Author: deepakkumar1984

Examples/SuperchargedArray.Testing/BasicExample.cs

@@ -52,10 +52,10 @@ public static void RunArraySimplified()
         {
             var K = Global.OP;
             //Create an array with values
-            NDArray a = new float[,] { { 1, 2 }, { 3, 4 }, { 5, 6 } };
+            SuperArray a = new float[,] { { 1, 2 }, { 3, 4 }, { 5, 6 } };
 
             //Create a array with constant value 2
-            NDArray b = new float[3, 2];
+            SuperArray b = new float[3, 2];
             //NDArray b = new NDArray(3, 2);
             b.Fill(2);
 

Examples/SuperchargedArray.Testing/PerfTest.cs

@@ -12,7 +12,7 @@ public class PerfTest
     {
         public void Run()
         {
-            int count = 10000000;
+            int count = 30000000;
             Random rnd = new Random();
 
             //Create variable A with random values
@@ -60,7 +60,7 @@ public void RunStandardLoop(int count, Array a, Array b)
             Console.WriteLine(".NET For Loop Time (in ms): " + sw.ElapsedMilliseconds);
         }
 
-        public void RunArrayDefault(int count, NDArray a, NDArray b, int cpu)
+        public void RunArrayDefault(int count, SuperArray a, SuperArray b, int cpu)
         {
             Global.UseDefault(cpu);
             Stopwatch sw = new Stopwatch();
@@ -74,7 +74,7 @@ public void RunArrayDefault(int count, NDArray a, NDArray b, int cpu)
             Console.WriteLine("With Parallel Thread Time (in ms): {1}", cpu, sw.ElapsedMilliseconds);
         }
 
-        public void RunArrayAccelerated(int count, NDArray a, NDArray b, int deviceid)
+        public void RunArrayAccelerated(int count, SuperArray a, SuperArray b, int deviceid)
         {
             try
             {
@@ -92,7 +92,7 @@ public void RunArrayAccelerated(int count, NDArray a, NDArray b, int deviceid)
             }
             catch(Exception ex)
             {
-                //Console.WriteLine("Error: " + ex.Message);
+                Console.WriteLine("Error: " + ex.Message);
             }
         }
     }

Examples/SuperchargedArray.Testing/SuperchargedArray.Testing.csproj

@@ -2,7 +2,7 @@
 
     <PropertyGroup>
         <OutputType>Exe</OutputType>
-        <TargetFramework>netcoreapp2.2</TargetFramework>
+        <TargetFramework>netcoreapp2.1</TargetFramework>
         <Platforms>AnyCPU;x64</Platforms>
     </PropertyGroup>
 

README.md

@@ -1,4 +1,4 @@
-# System.ArrayExtension
+# SuperchargedArray
 
 The .NET default built-in System.Array is very limited in terms of processing and usability. Here is the extended version of the Array with accelerated speed to execute operations on almost any hardware supporting OpenCL like Intel CPU, NVIDIA, AMD, Intel GPU, FPGA etc.
 
@@ -48,15 +48,15 @@ for (int i = 0; i < r.GetLength(0); i++)
 }
 ```
 
-A super simplified version with System.ArrayExtension
+A super simplified version with SuperchargedArray
 ```csharp
 var K = Global.OP;
 //Create an array with values
-NDArray a = new float[,] { { 1, 2 }, { 3, 4 }, { 5, 6 } };
+SuperArray a = new float[,] { { 1, 2 }, { 3, 4 }, { 5, 6 } };
 
 //Create a array with constant value 2
-NDArray b = new float[3, 2];
-//NDArray b = new NDArray(3, 2);
+SuperArray b = new float[3, 2];
+//SuperArray b = new SuperArray(3, 2);
 b.Fill(2);
 
 //Perform Math operation on the array: 2A + Log(B) + Exp(A)
@@ -68,7 +68,7 @@ r.Print();
 
 
 ## Accelerated Version
-System.ArrayExtension is not all about simplicity, it is well defined to execute the code on special hardwares like Intel GPU, NVIDIA, AMD cards. Below is the simple performance test done to process 100 million executions. And see the result:
+SuperchargedArray is not all about simplicity, it is well defined to execute the code on special hardwares like Intel GPU, NVIDIA, AMD cards. Below is the simple performance test done to process 100 million executions. And see the result:
 
 ```csharp
 public void Run()
@@ -121,29 +121,29 @@ public void RunStandardLoop(int count, Array a, Array b)
     Console.WriteLine(".NET For Loop Time (in ms): " + sw.ElapsedMilliseconds);
 }
 
-public void RunArrayDefault(int count, NDArray a, NDArray b, int cpu)
+public void RunArrayDefault(int count, SuperArray a, SuperArray b, int cpu)
 {
-    System.ArrayExtension.Accelerated.Global.UseDefault(cpu);
+    SuperchargedArray.Accelerated.Global.UseDefault(cpu);
     Stopwatch sw = new Stopwatch();
     sw.Start();
 
-    var K = System.ArrayExtension.Global.OP;
+    var K = SuperchargedArray.Global.OP;
 
     var r = K.Trunc(a * K.Sin(b) + K.Cos(a) * K.Exp(b));
     sw.Stop();
 
     Console.WriteLine("With Parallel Thread Time (in ms): {1}", cpu, sw.ElapsedMilliseconds);
 }
 
-public void RunArrayAccelerated(int count, NDArray a, NDArray b, int deviceid)
+public void RunArrayAccelerated(int count, SuperArray a, SuperArray b, int deviceid)
 {
     try
     {
         Accelerator.UseDevice(deviceid);
         Stopwatch sw = new Stopwatch();
         sw.Start();
 
-        var K = System.ArrayExtension.Accelerated.Global.OP;
+        var K = SuperchargedArray.Accelerated.Global.OP;
 
         var r = K.Trunc(a * K.Sin(b) + K.Cos(a) * K.Exp(b));
         sw.Stop();
@@ -160,19 +160,21 @@ public void RunArrayAccelerated(int count, NDArray a, NDArray b, int deviceid)
 
 ### Execution Result:
 
+Test result to execute math ops y = trunc(a * Sin(b) + cos(a) * exp(b)) for 100 million elements
+
 .NET For Loop Time (in ms): 22317
 
-Selected device: GeForce GTX 1080 Ti, With Accelerator (in ms): 8741
+GeForce GTX 1080 Ti                    : 8741 ms
 
-Selected device: Intel(R) UHD Graphics 630, With Accelerator (in ms): 6723
+Intel(R) UHD Graphics 630              : 6723 ms
 
-Selected device: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz, With Accelerator (in ms): 7380
+Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz: 7380 ms
 
-Selecting 33% of CPU processing, With Parallel Thread Time (in ms): 14332
+33% of CPU parallel processing                  : 14332 ms
 
-Selecting 66% of CPU processing, With Parallel Thread Time (in ms): 7716
+66% of CPU parallel processing                  : 7716 ms
 
-Selecting 100% of CPU processing, With Parallel Thread Time (in ms): 6420
+100% of CPU parallel processing:				: 6420 ms
 
 The .NET standard loop takes about 22 seconds, whereas with 100% CPU parallel approach using ArrayExtension it finishes off in 6 sec which is one-fourth of time taken. But for long running process using full CPU is not ideal. With ArrayExtension.Accelerated, which internally uses SIMD (Single Instruction Multiple Data), you can achive greater speed without using 100% of the hardware. But you can always fine-tune to use 100% of the hardware and achieve ultimate speed.
 

SuperchargedArray.sln

@@ -1,14 +1,21 @@
 
 Microsoft Visual Studio Solution File, Format Version 12.00
-# Visual Studio Version 16
-VisualStudioVersion = 16.0.28922.388
+# Visual Studio 15
+VisualStudioVersion = 15.0.28307.489
 MinimumVisualStudioVersion = 10.0.40219.1
 Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "examples", "examples", "{D6D26AE3-3049-476D-8014-2256876294B8}"
 EndProject
 Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "SuperchargedArray", "SuperchargedArray\SuperchargedArray.csproj", "{450BC48F-70F8-4E72-A837-CC4BBC85ED11}"
 EndProject
 Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "SuperchargedArray.Testing", "Examples\SuperchargedArray.Testing\SuperchargedArray.Testing.csproj", "{B87AA65D-77BF-47A8-B3F9-AC4ED142B89B}"
 EndProject
+Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "__", "__", "{C707FDE0-1468-4C6B-BB05-97F7243FA3DA}"
+	ProjectSection(SolutionItems) = preProject
+		.gitignore = .gitignore
+		LICENSE = LICENSE
+		README.md = README.md
+	EndProjectSection
+EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Any CPU = Debug|Any CPU

SuperchargedArray/Accelerated/Accelerator.cs

@@ -1,13 +1,35 @@
-using SuperchargedArray.Accelerated.OpenCL;
-using SuperchargedArray.Accelerated.OpenCL.Bindings;
-using System;
-using System.Collections.Generic;
-using System.Diagnostics;
-using System.IO;
-using System.Linq;
-
+/*
+MIT License
+
+Copyright (c) 2019 Tech Quantum
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+ 
+*/
 namespace SuperchargedArray.Accelerated
 {
+    using SuperchargedArray.Accelerated.OpenCL;
+    using System;
+    using System.Collections.Generic;
+    using System.IO;
+    using System.Linq;
+
     /// <summary>
     /// Accelerator is a high level API for building and executing built-in or your own kernels. 
     /// </summary>
@@ -165,12 +187,12 @@ public static void LoadKernals(string folderWithSourceCodes, string filter = "*.
         /// <param name="inputs">The inputs.</param>
         /// <param name="outputShape">The output shape.</param>
         /// <returns></returns>
-        public static NDArray Execute(string functionName, object[] inputs, long[] outputShape = null, int? returnResult = null)
+        public static SuperArray Execute(string functionName, object[] inputs, long[] outputShape = null, int? returnResult = null)
         {
             DType dType = DType.Single;
             if (outputShape == null)
             {
-                var ndobject = (NDArray)inputs.FirstOrDefault(x => (x.GetType() == typeof(NDArray)));
+                var ndobject = (SuperArray)inputs.FirstOrDefault(x => (x.GetType() == typeof(SuperArray)));
                 if (ndobject == null)
                 {
                     outputShape = new long[] { 1 };
@@ -189,16 +211,16 @@ public static NDArray Execute(string functionName, object[] inputs, long[] outpu
             else if (dType == DType.Double)
                 resultArray = ExecuteInternalArray<double>(functionName, inputs, dType, returnResult);
 
-            NDArray result = null;
+            SuperArray result = null;
 
             if (!returnResult.HasValue)
             {
-                result = new NDArray(outputShape, dType);
+                result = new SuperArray(outputShape, dType);
                 result.LoadFrom(resultArray);
             }
             else
             {
-                result = (NDArray)inputs[returnResult.Value];
+                result = (SuperArray)inputs[returnResult.Value];
                 result.LoadFrom(resultArray);
             }
 
@@ -269,7 +291,7 @@ private static Array ExecuteInternalArray<TSource>(string functionName, object[]
 
             try
             {
-                var ndobject = (NDArray)inputs.FirstOrDefault(x => (x.GetType() == typeof(NDArray)));
+                var ndobject = (SuperArray)inputs.FirstOrDefault(x => (x.GetType() == typeof(SuperArray)));
 
                 long length = ndobject != null ? ndobject.Elements : 1;
                 Array resultArray = null;
@@ -304,14 +326,14 @@ private static ComputeBuffer<TSource> BuildKernelArguments<TSource>(object[] inp
 
             foreach (var item in inputs)
             {
-                if (item.GetType() == typeof(NDArray))
+                if (item.GetType() == typeof(SuperArray))
                     if (returnResult.HasValue && returnResult.Value == i)
                     {
-                        result = new ComputeBuffer<TSource>(_context, ComputeMemoryFlags.ReadWrite | ComputeMemoryFlags.CopyHostPointer, ((NDArray)item).Data<TSource>());
+                        result = new ComputeBuffer<TSource>(_context, ComputeMemoryFlags.ReadWrite | ComputeMemoryFlags.CopyHostPointer, ((SuperArray)item).Data<TSource>());
                         kernel.SetMemoryArgument(i, result);
                     }
                     else
-                        kernel.SetMemoryArgument(i, new ComputeBuffer<TSource>(_context, ComputeMemoryFlags.ReadWrite | ComputeMemoryFlags.CopyHostPointer, ((NDArray)item).Data<TSource>()));
+                        kernel.SetMemoryArgument(i, new ComputeBuffer<TSource>(_context, ComputeMemoryFlags.ReadWrite | ComputeMemoryFlags.CopyHostPointer, ((SuperArray)item).Data<TSource>()));
                 else if (item.GetType().IsPrimitive)
                     kernel.SetValueArgument(i, (float)item);