clVector.DotProd Method

Overload List

#	Signature	Description
1	function DotProd(const Vec: TOpenCLMtxVec): Double;	Scalar product of two real arrays.
	└ indexed: function DotProd(const Vec: TOpenCLMtxVec; VecIndex: Integer; Index: Integer; Len: Integer): Double;
2	procedure DotProd(DstIndex: Integer; const Vec1: TOpenCLMtxVec; const Vec2: TOpenCLMtxVec; const Buffer: TOpenCLMtxVec);	Scalar product of two real arrays.
3	procedure DotProd(DstIndex: Integer; const Vec1: TOpenCLMtxVec; const Vec2: TOpenCLMtxVec; ConjVec: Boolean; const Buffer: TOpenCLMtxVec);	Scalar product of two real or complex arrays.

Overload 1: function DotProd(const Vec: TOpenCLMtxVec): Double;

Scalar product of two real arrays.

#	Name	Type	Description
1	Vec	TOpenCLMtxVec

Returns: Double

Remarks:

Calculates the dot product (scalar value) of the calling object and Vec object and returns a real scalar value. The dot product is defined by the equation:

\text{V}\cdot \text{V} = \sum _{k=0} ^{\text{Length}-1}\text{V}[k]\cdot \text{V}[k]

Both objects must be of equal size. If they are not, the method will return the dot product of the largest sub-array.

Examples

var a,b: clVector;
prod: double;
begin
    a.CopyFromAray(TDoubleArray.Create(1,2,3,4));
    b.CopyFromAray(TDoubleArray.Create(5,6,7,8));
    prod := a.DotProd(b); // = 1*5 + 2*6 + * 3*7 + 4*8
end;

See Also: clVector.DotProdc

Indexed: function DotProd(const Vec: TOpenCLMtxVec; VecIndex: Integer; Index: Integer; Len: Integer): Double;

Returns the scalar product between Vec elements [VecIndex]..[VecIndex+Len-1] and calling object elements [Index]..[Index+Len-1].

#	Name	Type	Description
1	Vec	TOpenCLMtxVec
2	VecIndex	Integer
3	Index	Integer	start index
4	Len	Integer	element count

Returns: Double

Remarks:

An exception is raised if Vec and calling object clVector.Complex property is True. An exception is raised if array borders are overrun or underrun.

Examples

var a,b: TOpenCLVector;
ac: TDoubleArray;
prod: double;
begin
    CreateIt(a,b,c);
    try
        a.CopyFromAray(TDoubleArray.Create(1,2,3,4));
        b.CopyFromAray(TDoubleArray.Create(5,6,7,8));
        c.Size(3);
        c.DotProd(0, a,b); // c[0] = 1*5 + 2*6 + * 3*7 + 4*8
        c.DotProd(1, a,b); // c[1] = 1*5 + 2*6 + * 3*7 + 4*8
        c.DotProd(2, a,b); // c[2] = 1*5 + 2*6 + * 3*7 + 4*8
        c.CopyToArray(ac); //ac = [70, 70, 70]
    finally
        FreeIt(a,b,c);
    end;
end;

Overload 2: procedure DotProd(DstIndex: Integer; const Vec1: TOpenCLMtxVec; const Vec2: TOpenCLMtxVec; const Buffer: TOpenCLMtxVec);

Scalar product of two real arrays.

#	Name	Type	Description
1	DstIndex	Integer	destination start index
2	Vec1	TOpenCLMtxVec
3	Vec2	TOpenCLMtxVec
4	Buffer	TOpenCLMtxVec

Result: stored in self (calling object)

Remarks:

Calculates the dot product (scalar value) of the calling object and Vec object and returns a real scalar value. The result is stored at specified index in the calling vector in the GPU memory. This variant of the function is non-blocking (faster), because the result does not have to be copied to the CPU memory to be used.

The dot product is defined by the equation:

\text{V}\cdot \text{V} = \sum _{k=0} ^{\text{Length}-1}\text{V}[k]\cdot \text{V}[k]

Both objects must be of equal size. If they are not, the method will return the dot product of the largest sub-array.

See Also: clVector.DotProdc

Overload 3: procedure DotProd(DstIndex: Integer; const Vec1: TOpenCLMtxVec; const Vec2: TOpenCLMtxVec; ConjVec: Boolean; const Buffer: TOpenCLMtxVec);

Scalar product of two real or complex arrays.

#	Name	Type	Description
1	DstIndex	Integer	destination start index
2	Vec1	TOpenCLMtxVec
3	Vec2	TOpenCLMtxVec
4	ConjVec	Boolean
5	Buffer	TOpenCLMtxVec

Result: stored in self (calling object)

Remarks:

Calculates the dot product (scalar value) of the Vec1 and Vec2 stores the result in calling vector at position DstIndex. ConjVec parameter is ignored if data is real. If ConjVec is true, the function computes: result = Vec1*conj(Vec2)

The dot product is defined by the equation:

\text{V}\cdot \text{V} = \sum _{k=0} ^{\text{Length}-1}\text{V}[k]\cdot \text{V}[k]

Both objects must be of equal size. If they are not, the method will return the dot product of the largest sub-array.

Examples

var a,b: clMatrix;
prod: double;
begin
    a.Size(2,2, clFloat, false);
    b.Size(2,2, clFloat, false);
    a.CopyFromAray(TDoubleArray.Create(1,2,3,4));
    b.CopyFromAray(TDoubleArray.Create(5,6,7,8));
    prod := a.DotProd(b); // = 1*5 + 2*6 + * 3*7 + 4*8
end;