diff --git a/C#/Installation script.iss b/C#/Installation script.iss
index 4ba41b2..753032f 100644
--- a/C#/Installation script.iss	
+++ b/C#/Installation script.iss	
@@ -3,7 +3,7 @@ AppName=ClooWrapperVBA
 AppVerName=ClooWrapperVBA
 DefaultDirName={pf}\ClooWrapperVBA
 DefaultGroupName=ClooWrapperVBA
-Compression=lzma
+Compression=zip
 SolidCompression=yes
 SourceDir=.\
 PrivilegesRequired=poweruser
@@ -22,9 +22,11 @@ Name: "{app}\demo\cl"; Permissions: everyone-full
 Source: bin\ClooWrapperVBA.dll; DestDir: {app}; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;  
 Source: bin\ClooWrapperVBA_x64.dll; DestDir: {app}; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
 Source: bin\Cloo.dll; DestDir: {app}; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
-Source: ..\Excel\OpenCl v0.05.xlsm; DestDir: {app}\demo; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
-Source: ..\Excel\cl\Performance.cl; DestDir: {app}\demo\cl; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
-Source: ..\Excel\cl\MatrixMultiplication.cl; DestDir: {app}\demo\cl; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
+Source: ..\Excel\OpenCl example.xlsm; DestDir: {app}\demo; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
+Source: ..\Excel\cl\FloatPerformance.cl; DestDir: {app}\demo\cl; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
+Source: ..\Excel\cl\DoublePerformance.cl; DestDir: {app}\demo\cl; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
+Source: ..\Excel\cl\FloatMatrixMultiplication.cl; DestDir: {app}\demo\cl; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
+Source: ..\Excel\cl\DoubleMatrixMultiplication.cl; DestDir: {app}\demo\cl; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
 Source: ..\Excel\Configuration.vbs; DestDir: {app}\demo; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
 Source: bin\register.bat; DestDir: {app}; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
 Source: bin\unregister.bat; DestDir: {app}; Flags: ignoreversion recursesubdirs overwritereadonly; Permissions: everyone-full;
diff --git a/Excel/Asynchronous.bas b/Excel/Asynchronous.bas
index d565543..1f323bb 100644
--- a/Excel/Asynchronous.bas
+++ b/Excel/Asynchronous.bas
@@ -10,7 +10,7 @@ Const THREAD_PRIORITY = 0
 Dim progDevices As Collection
 Dim currentTaskId&()
 
-Dim vecResp#()
+Dim vecResp!()
 Dim wsAsynchronous As Worksheet
 Dim globalWorkSize&(1), localWorkSize&(), globalWorkOffset&()
 Dim logLine%
@@ -26,7 +26,7 @@ Sub MainLoop()
     While Not allTasks_Completed
         For i = 1 To progDevices.Count
             If progDevices.Item(i).ProgramDevice.ExecutionCompleted Then
-                result = progDevices.Item(i).ProgramDevice.GetMemoryArgument_Double(0, vecResp)    ' Extract the results and do something with received data here.
+                result = progDevices.Item(i).ProgramDevice.GetMemoryArgument_Single(0, vecResp)    ' Extract the results and do something with received data here.
                 
                 wsAsynchronous.Cells(logLine, 1) = "Task " & currentTaskId(i) & ", " & progDevices.Item(i).ProgramDevice.deviceType & _
                     progDevices.Item(i).DeviceId & ": completed"
@@ -37,7 +37,7 @@ Sub MainLoop()
                 ' Start new task
                 If startedTasks < MAX_TASKS Then
                     ReDim vecResp(UBound(vecResp))  ' Erase output vector.
-                    result = progDevices.Item(i).ProgramDevice.SetMemoryArgument_Double(0, vecResp)
+                    result = progDevices.Item(i).ProgramDevice.SetMemoryArgument_Single(0, vecResp)
                     
                     ' If you want to use callbacks, than use function below
                     ' "CPU_Task_Completed" is a function that will obtain the callback.
@@ -83,14 +83,14 @@ Sub MainLoop()
 End Sub
 
 Sub RunAsynchronous()
-    Dim vecM1#(), vecM2#()
+    Dim vecM1!(), vecM2!()
     Dim vecQ&(1)
     Dim i&, j&, p&, q&, r&, nRows&
     Dim buildLogs$, sources$
     
     Set wsAsynchronous = ThisWorkbook.Worksheets("Asynchronous")
     
-    Open Application.ActiveWorkbook.Path & "\cl\MatrixMultiplication.cl" For Binary As #1
+    Open Application.ActiveWorkbook.Path & "\cl\FloatMatrixMultiplication.cl" For Binary As #1
     sources = Space$(LOF(1))
     Get #1, , sources
     Close #1
@@ -133,10 +133,10 @@ Sub RunAsynchronous()
     
     ReDim currentTaskId(progDevices.Count)
     For i = 1 To progDevices.Count
-        result = progDevices.Item(i).ProgramDevice.CreateKernel("DoubleMatrixMult")
-        result = progDevices.Item(i).ProgramDevice.SetMemoryArgument_Double(0, vecResp)
-        result = progDevices.Item(i).ProgramDevice.SetMemoryArgument_Double(1, vecM1)
-        result = progDevices.Item(i).ProgramDevice.SetMemoryArgument_Double(2, vecM2)
+        result = progDevices.Item(i).ProgramDevice.CreateKernel("FloatMatrixMult")
+        result = progDevices.Item(i).ProgramDevice.SetMemoryArgument_Single(0, vecResp)
+        result = progDevices.Item(i).ProgramDevice.SetMemoryArgument_Single(1, vecM1)
+        result = progDevices.Item(i).ProgramDevice.SetMemoryArgument_Single(2, vecM2)
         result = progDevices.Item(i).ProgramDevice.SetMemoryArgument_Long(3, vecQ)
     Next i
     
diff --git a/Excel/HelloWorld.bas b/Excel/HelloWorld.bas
index b2ec9bd..84c3eaf 100644
--- a/Excel/HelloWorld.bas
+++ b/Excel/HelloWorld.bas
@@ -6,10 +6,10 @@ Sub HelloWorld()
     Dim nRows%, currentRow&, nPlatforms&, nDevices&, i&, j&, result As Boolean
     Dim deviceType$, platformName$, platformVendor$, platformVersion$, deviceVendor$, deviceVersion$, driverVersion$, openCLCVersionString$
     Dim maxComputeUnits&, globalMemorySize#, maxClockFrequency#, maxMemoryAllocationSize#, deviceName$, sources$, cpuCounter&, gpuCounter&
-    Dim buildLogs$, platformId&, deviceId&, errorString$
+    Dim buildLogs$, platformId&, DeviceId&, errorString$
     Dim deviceAvailable As Boolean, compilerAvailable As Boolean
-    Dim m1#(1, 1), m2#(1, 1), vecM1#(), vecM2#(), vecQ&(0), vecResp#(3), globalWorkOffset&(), globalWorkSize&(1), localWorkSize&()
-    Dim p&, q&, r&, resp#()
+    Dim m1!(1, 1), m2!(1, 1), vecM1!(), vecM2!(), vecQ&(0), vecResp!(3), globalWorkOffset&(), globalWorkSize&(1), localWorkSize&()
+    Dim p&, q&, r&, resp!()
     
     Dim clooConfiguration As New ClooWrapperVBA.Configuration
     Dim progDevice As ClooWrapperVBA.ProgramDevice
@@ -76,7 +76,7 @@ Sub HelloWorld()
     
     ' Multiplication of two matrices.
     ' Read the OpenCL sources.
-    Open Application.ActiveWorkbook.Path & "\cl\MatrixMultiplication.cl" For Binary As #1
+    Open Application.ActiveWorkbook.Path & "\cl\FloatMatrixMultiplication.cl" For Binary As #1
     sources = Space$(LOF(1))
     Get #1, , sources
     Close #1
@@ -87,13 +87,13 @@ Sub HelloWorld()
         result = clooConfiguration.SetPlatform(platformId)
         cpuCounter = 0
         gpuCounter = 0
-        For deviceId = 0 To clooConfiguration.Platform.Devices - 1
-            result = clooConfiguration.Platform.SetDevice(deviceId)
+        For DeviceId = 0 To clooConfiguration.Platform.Devices - 1
+            result = clooConfiguration.Platform.SetDevice(DeviceId)
             
             If clooConfiguration.Platform.device.compilerAvailable Then
                 If clooConfiguration.Platform.device.deviceType = "CPU" Then
                     Set progDevice = New ClooWrapperVBA.ProgramDevice
-                    result = progDevice.Build(sources, "", platformId, deviceId, cpuCounter, buildLogs)
+                    result = progDevice.Build(sources, "", platformId, DeviceId, cpuCounter, buildLogs)
                     If result Then
                         Exit Do
                     Else
@@ -102,7 +102,7 @@ Sub HelloWorld()
                 End If
                 If clooConfiguration.Platform.device.deviceType = "GPU" Then
                     Set progDevice = New ClooWrapperVBA.ProgramDevice
-                    result = progDevice.Build(sources, "", platformId, deviceId, gpuCounter, buildLogs)
+                    result = progDevice.Build(sources, "", platformId, DeviceId, gpuCounter, buildLogs)
                     gpuCounter = gpuCounter + 1
                     If result Then
                         Exit Do
@@ -111,12 +111,12 @@ Sub HelloWorld()
                     End If
                 End If
             End If
-        Next deviceId
+        Next DeviceId
         platformId = platformId + 1
     Loop
     
     errorString = progDevice.errorString
-    result = progDevice.CreateKernel("DoubleMatrixMult")
+    result = progDevice.CreateKernel("FloatMatrixMult")
     
     ' Initialization of arrays:
     p = 2: q = 2: r = 2
@@ -125,18 +125,18 @@ Sub HelloWorld()
             m1(i, j) = wsHelloWorld.Cells(i + 1, j + 7)
         Next j
     Next i
-    vecM1 = MatrixToVector(m1, p, q)
+    vecM1 = MatrixToVectorSingle(m1, p, q)
     For i = 0 To q - 1
         For j = 0 To r - 1
             m2(i, j) = wsHelloWorld.Cells(i + 3, j + 7)
         Next j
     Next i
-    vecM2 = MatrixToVector(m2, q, r)
+    vecM2 = MatrixToVectorSingle(m2, q, r)
     vecQ(0) = q
     
-    result = progDevice.SetMemoryArgument_Double(0, vecResp)
-    result = progDevice.SetMemoryArgument_Double(1, vecM1)
-    result = progDevice.SetMemoryArgument_Double(2, vecM2)
+    result = progDevice.SetMemoryArgument_Single(0, vecResp)
+    result = progDevice.SetMemoryArgument_Single(1, vecM1)
+    result = progDevice.SetMemoryArgument_Single(2, vecM2)
     result = progDevice.SetMemoryArgument_Long(3, vecQ)
     
     globalWorkSize(0) = p
@@ -144,9 +144,9 @@ Sub HelloWorld()
     
     result = progDevice.ExecuteSync(globalWorkOffset, globalWorkSize, localWorkSize)
     
-    result = progDevice.GetMemoryArgument_Double(0, vecResp)
+    result = progDevice.GetMemoryArgument_Single(0, vecResp)
     
-    resp = VectorToMatrix(vecResp, p, r)
+    resp = VectorToMatrixSingle(vecResp, p, r)
     
     For i = 0 To p - 1
         For j = 0 To r - 1
diff --git a/Excel/Helpers.bas b/Excel/Helpers.bas
index 83e3420..7d66959 100644
--- a/Excel/Helpers.bas
+++ b/Excel/Helpers.bas
@@ -37,12 +37,12 @@ Function CreateDeviceCollection(sources$)
             
             If clooConfiguration.Platform.device.deviceType = "CPU" Then
                 result = progDevice.ProgramDevice.Build(sources, "", i - 1, j - 1, cpuCounter, buildLogs)
-                progDevice.deviceId = cpuCounter
+                progDevice.DeviceId = cpuCounter
                 progDevice.deviceType = "CPU"
                 If result = True Then cpuCounter = cpuCounter + 1
             ElseIf clooConfiguration.Platform.device.deviceType = "GPU" Then
                 result = progDevice.ProgramDevice.Build(sources, "", i - 1, j - 1, gpuCounter, buildLogs)
-                progDevice.deviceId = gpuCounter
+                progDevice.DeviceId = gpuCounter
                 progDevice.deviceType = "GPU"
                 If result = True Then gpuCounter = gpuCounter + 1
             Else
@@ -62,7 +62,37 @@ Function CreateDeviceCollection(sources$)
     End If
 End Function
 
-Function MatrixToVector(m() As Double, maxi As Long, maxj As Long) As Double()
+Function MatrixToVectorSingle(m() As Single, maxi As Long, maxj As Long) As Single()
+    Dim v() As Single
+    Dim i&, j&
+    
+    ReDim v(maxi * maxj - 1)
+    
+    For i = 0 To maxi - 1
+        For j = 0 To maxj - 1
+            v(i + maxi * j) = m(i, j)
+        Next j
+    Next i
+    
+    MatrixToVectorSingle = v
+End Function
+
+Function VectorToMatrixSingle(v() As Single, maxi As Long, maxj As Long) As Single()
+    Dim i&, j&
+    Dim m() As Single
+    
+    ReDim m(maxi - 1, maxj - 1)
+    
+    For i = 0 To maxi - 1
+        For j = 0 To maxj - 1
+            m(i, j) = v(i + maxi * j)
+        Next j
+    Next i
+    
+    VectorToMatrixSingle = m
+End Function
+
+Function MatrixToVectorDouble(m() As Double, maxi As Long, maxj As Long) As Double()
     Dim v() As Double
     Dim i&, j&
     
@@ -74,10 +104,10 @@ Function MatrixToVector(m() As Double, maxi As Long, maxj As Long) As Double()
         Next j
     Next i
     
-    MatrixToVector = v
+    MatrixToVectorDouble = v
 End Function
 
-Function VectorToMatrix(v() As Double, maxi As Long, maxj As Long) As Double()
+Function VectorToMatrixDouble(v() As Double, maxi As Long, maxj As Long) As Double()
     Dim i&, j&
     Dim m() As Double
     
@@ -89,5 +119,5 @@ Function VectorToMatrix(v() As Double, maxi As Long, maxj As Long) As Double()
         Next j
     Next i
     
-    VectorToMatrix = m
+    VectorToMatrixDouble = m
 End Function
diff --git a/Excel/OpenCl example.xlsm b/Excel/OpenCl example.xlsm
index 2bf08ac..447c252 100644
Binary files a/Excel/OpenCl example.xlsm and b/Excel/OpenCl example.xlsm differ
diff --git a/Excel/Performance.bas b/Excel/Performance.bas
index 8eeba67..e6766b9 100644
--- a/Excel/Performance.bas
+++ b/Excel/Performance.bas
@@ -1,13 +1,13 @@
 Attribute VB_Name = "Performance"
 Option Explicit
 
-Private Const ARRAY_SIZE = 1200
+Private Const ARRAY_SIZE = 1000
 
 Sub VBA_PerformanceTest()
     Dim wsPerformanceTest As Worksheet
-    Dim m1#(), m2#(), vecM1#(), vecM2#(), vecResp#(), resultVba#(), vecQ&(0)
-    Dim x1#(0), x2#(0), res#(0)
-    Dim finalResults#()
+    Dim m1!(), m2!(), vecM1!(), vecM2!(), vecResp!(), resultVba!(), vecQ&(0)
+    Dim x1!(0), x2!(0), res!(0)
+    Dim finalResults!()
     Dim i&, j&, k&, p&, q&, r&
     Dim buildLogs$, sources$, result As Boolean
     Dim cTime As New CTimer
@@ -31,17 +31,17 @@ Sub VBA_PerformanceTest()
     Randomize
     For i = 0 To p - 1
         For j = 0 To q - 1
-            m1(i, j) = (Rnd() - 0.5) * 10#
+            m1(i, j) = CInt((Rnd() - 0.5) * 100#)
         Next j
     Next i
     
     For i = 0 To q - 1
         For j = 0 To r - 1
-            m2(i, j) = (Rnd() - 0.5) * 10#
+            m2(i, j) = CInt((Rnd() - 0.5) * 100#)
         Next j
     Next i
-    vecM1 = MatrixToVector(m1, p, q)
-    vecM2 = MatrixToVector(m2, q, r)
+    vecM1 = MatrixToVectorSingle(m1, p, q)
+    vecM2 = MatrixToVectorSingle(m2, q, r)
     
     ' VBA matrix multiplication:
     cTime.StartCounter
@@ -54,7 +54,7 @@ Sub VBA_PerformanceTest()
     Next i
     wsPerformanceTest.Cells(2, 2) = cTime.TimeElapsed
     
-    Open Application.ActiveWorkbook.Path & "\cl\MatrixMultiplication.cl" For Binary As #1
+    Open Application.ActiveWorkbook.Path & "\cl\FloatMatrixMultiplication.cl" For Binary As #1
     sources = Space$(LOF(1))
     Get #1, , sources
     Close #1
@@ -71,15 +71,15 @@ Sub VBA_PerformanceTest()
         ' CPU calculations.
         Set programDevice_Performance = GetFirstDeviceOfType(progDevices, "CPU")
     
-        result = programDevice_Performance.CreateKernel("DoubleMatrixMult")
+        result = programDevice_Performance.CreateKernel("FloatMatrixMult")
     
         globalWorkSize(0) = p
         globalWorkSize(1) = r
         vecQ(0) = q
     
-        result = programDevice_Performance.SetMemoryArgument_Double(0, vecResp)
-        result = programDevice_Performance.SetMemoryArgument_Double(1, vecM1)
-        result = programDevice_Performance.SetMemoryArgument_Double(2, vecM2)
+        result = programDevice_Performance.SetMemoryArgument_Single(0, vecResp)
+        result = programDevice_Performance.SetMemoryArgument_Single(1, vecM1)
+        result = programDevice_Performance.SetMemoryArgument_Single(2, vecM2)
         result = programDevice_Performance.SetMemoryArgument_Long(3, vecQ)
     
         ' Start once to update cashes.
@@ -90,8 +90,8 @@ Sub VBA_PerformanceTest()
         result = programDevice_Performance.ExecuteSync(globalWorkOffset, globalWorkSize, localWorkSize)
         wsPerformanceTest.Cells(3, 2) = cTime.TimeElapsed
     
-        result = programDevice_Performance.GetMemoryArgument_Double(0, vecResp)
-        finalResults = VectorToMatrix(vecResp, p, r)
+        result = programDevice_Performance.GetMemoryArgument_Single(0, vecResp)
+        finalResults = VectorToMatrixSingle(vecResp, p, r)
     
         ' Comparison to VBA result.
         calcCorrect = True
@@ -119,16 +119,16 @@ Sub VBA_PerformanceTest()
     If Not (GetFirstDeviceOfType(progDevices, "GPU") Is Nothing) Then
         Set programDevice_Performance = GetFirstDeviceOfType(progDevices, "GPU")
         
-        result = programDevice_Performance.CreateKernel("DoubleMatrixMult")
+        result = programDevice_Performance.CreateKernel("FloatMatrixMult")
         
         globalWorkSize(0) = p
         globalWorkSize(1) = r
         vecQ(0) = q
         
         ReDim vecResp(p * r - 1)
-        result = programDevice_Performance.SetMemoryArgument_Double(0, vecResp)
-        result = programDevice_Performance.SetMemoryArgument_Double(1, vecM1)
-        result = programDevice_Performance.SetMemoryArgument_Double(2, vecM2)
+        result = programDevice_Performance.SetMemoryArgument_Single(0, vecResp)
+        result = programDevice_Performance.SetMemoryArgument_Single(1, vecM1)
+        result = programDevice_Performance.SetMemoryArgument_Single(2, vecM2)
         result = programDevice_Performance.SetMemoryArgument_Long(3, vecQ)
         
         ' Start once to update cashes.
@@ -139,8 +139,8 @@ Sub VBA_PerformanceTest()
         result = programDevice_Performance.ExecuteSync(globalWorkOffset, globalWorkSize, localWorkSize)
         wsPerformanceTest.Cells(4, 2) = cTime.TimeElapsed
         
-        result = programDevice_Performance.GetMemoryArgument_Double(0, vecResp)
-        finalResults = VectorToMatrix(vecResp, p, r)
+        result = programDevice_Performance.GetMemoryArgument_Single(0, vecResp)
+        finalResults = VectorToMatrixSingle(vecResp, p, r)
         
         ' Comparison to VBA result.
         calcCorrect = True
@@ -164,28 +164,30 @@ Sub VBA_PerformanceTest()
     End If
 End Sub
 
-Sub GpuCpu_SingleDouble_PerformanceTest()
+Sub GpuCpu_FloatDouble_PerformanceTest()
+    GpuCpu_Float_PerformanceTest
+    GpuCpu_Double_PerformanceTest
+End Sub
+
+Sub GpuCpu_Float_PerformanceTest()
     Dim wsPerformanceTest As Worksheet
-    Dim upper&, singles!(), doubles#(), aSingle!, aDouble#, i&
+    Dim upper&, singles!(), aSingle!, i&
     Dim sources$, result As Boolean
     Dim progDevices As Collection
     Dim programDevice_Performance As ClooWrapperVBA.ProgramDevice
     
     Set wsPerformanceTest = ThisWorkbook.Worksheets("Performance")
-    wsPerformanceTest.Range("E3:F4").ClearContents
+    wsPerformanceTest.Range("E3:E4").ClearContents
     
     upper = 10000000
     ReDim singles(upper)
-    ReDim doubles(upper)
     
     For i = 0 To upper - 1
         singles(i) = i
-        doubles(i) = i
     Next i
     aSingle = 2!
-    aDouble = 2#
     
-    Open Application.ActiveWorkbook.Path & "\cl\Performance.cl" For Binary As #1
+    Open Application.ActiveWorkbook.Path & "\cl\FloatPerformance.cl" For Binary As #1
     sources = Space$(LOF(1))
     Get #1, , sources
     Close #1
@@ -200,11 +202,9 @@ Sub GpuCpu_SingleDouble_PerformanceTest()
     
     If GetFirstDeviceOfType(progDevices, "GPU") Is Nothing Then
         wsPerformanceTest.Cells(4, 5) = CVErr(2042)
-        wsPerformanceTest.Cells(4, 6) = CVErr(2042)
     Else
         Set programDevice_Performance = GetFirstDeviceOfType(progDevices, "GPU")
-        wsPerformanceTest.Cells(4, 5) = GPU_PerformanceTest_Single(upper, singles, aSingle, programDevice_Performance)
-        wsPerformanceTest.Cells(4, 6) = GPU_PerformanceTest_Double(upper, doubles, aDouble, programDevice_Performance)
+        wsPerformanceTest.Cells(4, 5) = PerformanceTest_Single(upper, singles, aSingle, programDevice_Performance)
         
         result = programDevice_Performance.ReleaseMemObject(1)
         result = programDevice_Performance.ReleaseMemObject(0)
@@ -214,11 +214,9 @@ Sub GpuCpu_SingleDouble_PerformanceTest()
     
     If GetFirstDeviceOfType(progDevices, "CPU") Is Nothing Then
         wsPerformanceTest.Cells(3, 5) = CVErr(2042)
-        wsPerformanceTest.Cells(3, 6) = CVErr(2042)
     Else
         Set programDevice_Performance = GetFirstDeviceOfType(progDevices, "CPU")
-        wsPerformanceTest.Cells(3, 5) = CPU_PerformanceTest_Single(upper, singles, aSingle, programDevice_Performance)
-        wsPerformanceTest.Cells(3, 6) = CPU_PerformanceTest_Double(upper, doubles, aDouble, programDevice_Performance)
+        wsPerformanceTest.Cells(3, 5) = PerformanceTest_Single(upper, singles, aSingle, programDevice_Performance)
         
         result = programDevice_Performance.ReleaseMemObject(1)
         result = programDevice_Performance.ReleaseMemObject(0)
@@ -227,44 +225,76 @@ Sub GpuCpu_SingleDouble_PerformanceTest()
     End If
 End Sub
 
-' Single precision performance at GPU.
-Function GPU_PerformanceTest_Single(upper&, singles!(), aSingle!, programDevice_Performance As ClooWrapperVBA.ProgramDevice)
-    Dim buildLogs$, result As Boolean
+Sub GpuCpu_Double_PerformanceTest()
+    Dim wsPerformanceTest As Worksheet
+    Dim upper&, doubles#(), aDouble#, i&
+    Dim sources$, result As Boolean
+    Dim progDevices As Collection
+    Dim programDevice_Performance As ClooWrapperVBA.ProgramDevice
     
-    result = programDevice_Performance.CreateKernel("SinglePerformance")
+    Set wsPerformanceTest = ThisWorkbook.Worksheets("Performance")
+    wsPerformanceTest.Range("F3:F4").ClearContents
     
-    result = programDevice_Performance.SetMemoryArgument_Single(0, singles)
-    result = programDevice_Performance.SetValueArgument_Single(1, aSingle)
+    upper = 10000000
+    ReDim doubles(upper)
     
-    GPU_PerformanceTest_Single = PerformanceTestExecution(upper, programDevice_Performance)
-End Function
-
-' Single precision performance at CPU.
-Function CPU_PerformanceTest_Single(upper&, singles!(), aSingle!, programDevice_Performance As ClooWrapperVBA.ProgramDevice)
-    Dim buildLogs$, result As Boolean
+    For i = 0 To upper - 1
+        doubles(i) = i
+    Next i
+    aDouble = 2#
     
-    result = programDevice_Performance.CreateKernel("SinglePerformance")
+    Open Application.ActiveWorkbook.Path & "\cl\DoublePerformance.cl" For Binary As #1
+    sources = Space$(LOF(1))
+    Get #1, , sources
+    Close #1
     
-    result = programDevice_Performance.SetMemoryArgument_Single(0, singles)
-    result = programDevice_Performance.SetValueArgument_Single(1, aSingle)
+    ' Adding of all CPU and GPU devices to collection.
+    Set progDevices = CreateDeviceCollection(sources)
     
-    CPU_PerformanceTest_Single = PerformanceTestExecution(upper, programDevice_Performance)
-End Function
+    If progDevices Is Nothing Then
+        MsgBox ("No devices found! Something is wrong!")
+        Exit Sub
+    End If
+    
+    If GetFirstDeviceOfType(progDevices, "GPU") Is Nothing Then
+        wsPerformanceTest.Cells(4, 6) = CVErr(2042)
+    Else
+        Set programDevice_Performance = GetFirstDeviceOfType(progDevices, "GPU")
+        wsPerformanceTest.Cells(4, 6) = PerformanceTest_Double(upper, doubles, aDouble, programDevice_Performance)
+        
+        result = programDevice_Performance.ReleaseMemObject(1)
+        result = programDevice_Performance.ReleaseMemObject(0)
+        result = programDevice_Performance.ReleaseKernel
+        result = programDevice_Performance.ReleaseProgram
+    End If
+    
+    If GetFirstDeviceOfType(progDevices, "CPU") Is Nothing Then
+        wsPerformanceTest.Cells(3, 6) = CVErr(2042)
+    Else
+        Set programDevice_Performance = GetFirstDeviceOfType(progDevices, "CPU")
+        wsPerformanceTest.Cells(3, 6) = PerformanceTest_Double(upper, doubles, aDouble, programDevice_Performance)
+        
+        result = programDevice_Performance.ReleaseMemObject(1)
+        result = programDevice_Performance.ReleaseMemObject(0)
+        result = programDevice_Performance.ReleaseKernel
+        result = programDevice_Performance.ReleaseProgram
+    End If
+End Sub
 
-' Double precision performance at GPU.
-Function GPU_PerformanceTest_Double(upper&, doubles#(), aDouble#, programDevice_Performance As ClooWrapperVBA.ProgramDevice)
+' Single precision performance at CPU / GPU.
+Function PerformanceTest_Single(upper&, singles!(), aSingle!, programDevice_Performance As ClooWrapperVBA.ProgramDevice)
     Dim buildLogs$, result As Boolean
     
-    result = programDevice_Performance.CreateKernel("DoublePerformance")
+    result = programDevice_Performance.CreateKernel("FloatPerformance")
     
-    result = programDevice_Performance.SetMemoryArgument_Double(0, doubles)
-    result = programDevice_Performance.SetValueArgument_Double(1, aDouble)
+    result = programDevice_Performance.SetMemoryArgument_Single(0, singles)
+    result = programDevice_Performance.SetValueArgument_Single(1, aSingle)
     
-    GPU_PerformanceTest_Double = PerformanceTestExecution(upper, programDevice_Performance)
+    PerformanceTest_Single = PerformanceTestExecution(upper, programDevice_Performance)
 End Function
 
-' Double precision performance at CPU.
-Function CPU_PerformanceTest_Double(upper&, doubles#(), aDouble#, programDevice_Performance As ClooWrapperVBA.ProgramDevice)
+' Double precision performance at CPU / GPU.
+Function PerformanceTest_Double(upper&, doubles#(), aDouble#, programDevice_Performance As ClooWrapperVBA.ProgramDevice)
     Dim buildLogs$, result As Boolean
     
     result = programDevice_Performance.CreateKernel("DoublePerformance")
@@ -272,7 +302,7 @@ Function CPU_PerformanceTest_Double(upper&, doubles#(), aDouble#, programDevice_
     result = programDevice_Performance.SetMemoryArgument_Double(0, doubles)
     result = programDevice_Performance.SetValueArgument_Double(1, aDouble)
     
-    CPU_PerformanceTest_Double = PerformanceTestExecution(upper, programDevice_Performance)
+    PerformanceTest_Double = PerformanceTestExecution(upper, programDevice_Performance)
 End Function
 
 Function PerformanceTestExecution(upper&, programDevice_Performance)
@@ -305,9 +335,9 @@ Function PerformanceTestExecution(upper&, programDevice_Performance)
 End Function
 
 Sub Test_OneAfterAnother()
-    Dim m1#(), m2#(), vecM1#(), vecM2#(), vecResp#(), resultVba#(), vecQ&(0)
-    Dim x1#(0), x2#(0), res#(0)
-    Dim finalResults#()
+    Dim m1!(), m2!(), vecM1!(), vecM2!(), vecResp!(), resultVba!(), vecQ&(0)
+    Dim x1!(0), x2!(0), res!(0)
+    Dim finalResults!()
     Dim i&, j&, k&, p&, q&, r&
     Dim buildLogs$, sources$, result As Boolean
     Dim cTime As New CTimer
@@ -340,7 +370,7 @@ Sub Test_OneAfterAnother()
         Next j
     Next i
     
-    Open Application.ActiveWorkbook.Path & "\cl\MatrixMultiplication.cl" For Binary As #1
+    Open Application.ActiveWorkbook.Path & "\cl\FloatMatrixMultiplication.cl" For Binary As #1
     sources = Space$(LOF(1))
     Get #1, , sources
     Close #1
@@ -357,22 +387,22 @@ Sub Test_OneAfterAnother()
         ' CPU calculations.
         Set programDevice_Performance = GetFirstDeviceOfType(progDevices, "CPU")
     
-        result = programDevice_Performance.CreateKernel("DoubleMatrixMult")
+        result = programDevice_Performance.CreateKernel("FloatMatrixMult")
     
         globalWorkSize(0) = p
         globalWorkSize(1) = r
         vecQ(0) = q
     
-        result = programDevice_Performance.SetMemoryArgument_Double(0, vecResp)
-        result = programDevice_Performance.SetMemoryArgument_Double(1, vecM1)
-        result = programDevice_Performance.SetMemoryArgument_Double(2, vecM2)
+        result = programDevice_Performance.SetMemoryArgument_Single(0, vecResp)
+        result = programDevice_Performance.SetMemoryArgument_Single(1, vecM1)
+        result = programDevice_Performance.SetMemoryArgument_Single(2, vecM2)
         result = programDevice_Performance.SetMemoryArgument_Long(3, vecQ)
         
         result = programDevice_Performance.ExecuteSync(globalWorkOffset, globalWorkSize, localWorkSize)
         
-        result = programDevice_Performance.GetMemoryArgument_Double(0, vecResp)
+        result = programDevice_Performance.GetMemoryArgument_Single(0, vecResp)
         finalResults = VectorToMatrix(vecResp, p, r)
-    
+        
         ' Comparison to VBA result.
         calcCorrect = True
         For i = 0 To p - 1
@@ -383,12 +413,11 @@ Sub Test_OneAfterAnother()
             Next j
         Next i
         
-        
-        result = programDevice_Performance.SetMemoryArgument_Double(1, vecM2)
-        result = programDevice_Performance.SetMemoryArgument_Double(2, vecM1)
+        result = programDevice_Performance.SetMemoryArgument_Single(1, vecM2)
+        result = programDevice_Performance.SetMemoryArgument_Single(2, vecM1)
         
         result = programDevice_Performance.ExecuteSync(globalWorkOffset, globalWorkSize, localWorkSize)
-        result = programDevice_Performance.GetMemoryArgument_Double(0, vecResp)
+        result = programDevice_Performance.GetMemoryArgument_Single(0, vecResp)
         finalResults = VectorToMatrix(vecResp, p, r)
         
         ' VBA matrix multiplication:
diff --git a/Excel/cl/DoubleMatrixMultiplication.cl b/Excel/cl/DoubleMatrixMultiplication.cl
new file mode 100644
index 0000000..ebd9435
--- /dev/null
+++ b/Excel/cl/DoubleMatrixMultiplication.cl
@@ -0,0 +1,15 @@
+__kernel void
+DoubleMatrixMult(__global double* MResp, __global double* M1, __global double* M2, __global int* q)
+{
+    // Vector element index
+    int i = get_global_id(0);
+    int j = get_global_id(1);
+    int p = get_global_size(0);
+    int r = get_global_size(1);
+    MResp[i + p * j] = 0;
+    int QQ = q[0];
+    for (int k = 0; k < QQ; k++)
+    {
+        MResp[i + p * j] += M1[i + p * k] * M2[k + QQ * j];
+    }
+}
\ No newline at end of file
diff --git a/Excel/cl/Performance.cl b/Excel/cl/DoublePerformance.cl
similarity index 83%
rename from Excel/cl/Performance.cl
rename to Excel/cl/DoublePerformance.cl
index ac50347..fba72bc 100644
--- a/Excel/cl/Performance.cl
+++ b/Excel/cl/DoublePerformance.cl
@@ -15,16 +15,6 @@
 #define MAD_1024(x,y)	MAD_256(x, y);		MAD_256(x, y);		MAD_256(x, y);		MAD_256(x, y);
 #define MAD_4096(x,y)	MAD_1024(x, y);		MAD_1024(x, y);		MAD_1024(x, y);		MAD_1024(x, y);
 
-__kernel void SinglePerformance(__global float* ptr, float _A) {
-    float x = _A;
-    float y = (float)get_local_id(0);
-
-    MAD_1024(x, y);
-    MAD_1024(x, y);
-
-    ptr[get_global_id(0)] = y;
-}
-
 __kernel void DoublePerformance(__global double* ptr, double _A) {
 	double x = _A;
 	double y = (double)get_local_id(0);
diff --git a/Excel/cl/FloatMatrixMultiplication.cl b/Excel/cl/FloatMatrixMultiplication.cl
new file mode 100644
index 0000000..33638f9
--- /dev/null
+++ b/Excel/cl/FloatMatrixMultiplication.cl
@@ -0,0 +1,15 @@
+__kernel void
+FloatMatrixMult(__global float* MResp, __global float* M1, __global float* M2, __global int* q)
+{
+    // Vector element index
+    int i = get_global_id(0);
+    int j = get_global_id(1);
+    int p = get_global_size(0);
+    int r = get_global_size(1);
+    MResp[i + p * j] = 0;
+    int QQ = q[0];
+    for (int k = 0; k < QQ; k++)
+    {
+        MResp[i + p * j] += M1[i + p * k] * M2[k + QQ * j];
+    }
+}
\ No newline at end of file
diff --git a/Excel/cl/FloatPerformance.cl b/Excel/cl/FloatPerformance.cl
new file mode 100644
index 0000000..60b55f4
--- /dev/null
+++ b/Excel/cl/FloatPerformance.cl
@@ -0,0 +1,20 @@
+#undef MAD_4
+#undef MAD_16
+#undef MAD_64
+
+#define MAD_4(x, y)     x = y*x+y;			y = x*y+x;			x = y*x+y;			y = x*y+x;
+#define MAD_16(x, y)    MAD_4(x, y);        MAD_4(x, y);        MAD_4(x, y);        MAD_4(x, y);
+#define MAD_64(x, y)    MAD_16(x, y);       MAD_16(x, y);       MAD_16(x, y);       MAD_16(x, y);
+#define MAD_256(x,y)	MAD_64(x, y);		MAD_64(x, y);		MAD_64(x, y);		MAD_64(x, y);
+#define MAD_1024(x,y)	MAD_256(x, y);		MAD_256(x, y);		MAD_256(x, y);		MAD_256(x, y);
+#define MAD_4096(x,y)	MAD_1024(x, y);		MAD_1024(x, y);		MAD_1024(x, y);		MAD_1024(x, y);
+
+__kernel void FloatPerformance(__global float* ptr, float _A) {
+    float x = _A;
+    float y = (float)get_local_id(0);
+
+    MAD_1024(x, y);
+    MAD_1024(x, y);
+
+    ptr[get_global_id(0)] = y;
+}
\ No newline at end of file
diff --git a/Excel/cl/MatrixMultiplication.cl b/Excel/cl/MatrixMultiplication.cl
deleted file mode 100644
index 3ff6e81..0000000
--- a/Excel/cl/MatrixMultiplication.cl
+++ /dev/null
@@ -1,31 +0,0 @@
-__kernel void
-FloatMatrixMult(__global float* MResp, __global float* M1, __global float* M2, __global int* q)
-{
-    // Vector element index
-    int i = get_global_id(0);
-    int j = get_global_id(1);
-    int p = get_global_size(0);
-    int r = get_global_size(1);
-    MResp[i + p * j] = 0;
-    int QQ = q[0];
-    for (int k = 0; k < QQ; k++)
-    {
-        MResp[i + p * j] += M1[i + p * k] * M2[k + QQ * j];
-    }
-}
-
-__kernel void
-DoubleMatrixMult(__global double* MResp, __global double* M1, __global double* M2, __global int* q)
-{
-    // Vector element index
-    int i = get_global_id(0);
-    int j = get_global_id(1);
-    int p = get_global_size(0);
-    int r = get_global_size(1);
-    MResp[i + p * j] = 0;
-    int QQ = q[0];
-    for (int k = 0; k < QQ; k++)
-    {
-        MResp[i + p * j] += M1[i + p * k] * M2[k + QQ * j];
-    }
-}
\ No newline at end of file
diff --git a/Readme.md b/Readme.md
index c516bfb..a812c66 100644
--- a/Readme.md
+++ b/Readme.md
@@ -82,5 +82,4 @@ The current version can be downloaded as:
         - Solution: Get OpenCL.dll from other computer.
     - Reason: The GPGPU / CPU drivers are too old and not supported by OpenCL.
         - Update the drivers.
-    - Regarding Xeon processors, it is recommended that you install the latest drivers, as detailed in the article "[OpenCL™ Drivers and Runtimes for Intel® Architecture](https://www.codeproject.com/Articles/1230011/OpenCL-Drivers-and-Runtimes-for-Intel-Architectu)".
 <br>