OpenCLTemplate

Class to help editing OpenCL code

Constructor. Takes care of a Rich Text Box KeyUp event to paint things

RichTextBox to control

Help indentation?

Regular text color

Regular text font

Comments color

Comments font

List of string structures to mark

KeyUp event handler

Forces OpenCL RichTextBox to be updated

Stops updating text box while coloring text

Rich text box to lock

Restarts updating text box

Rich text box to unlock

Defines a structure of strings to mark

Description of string type

Color to use for this string type

Font to be used in this list of strings

List of strings of this type

Displays OpenCL related information

Constructor.

Required designer variable.

Clean up any resources being used.

true if managed resources should be disposed; otherwise, false.

Required method for Designer support - do not modify the contents of this method with the code editor.

The main entry point for the application.

OpenCL Helper Editor

Required designer variable.

Clean up any resources being used.

true if managed resources should be disposed; otherwise, false.

Required method for Designer support - do not modify the contents of this method with the code editor.

Constructor.

Button to test code

OpenCL calculations class

String to include to enable Double Precision calculations

Currently used acceleration

Initialization error

Sets CLCalc status to NotUsingCL

Initializes OpenCL and reads devices

Initializes OpenCL from an existing Context and reads devices

Existing context pointer Existing command queue pointer

Tries to execute actual code in device to check its availability.

Command queue number to check

Releases OpenCL resources

Last found error

List of available platforms

List of available devices

Gets string to include to enable Double Precision calculations

Gets acceleration type being used

Gets initialization error description

Last error. Throws exception if not set to success.

OpenCL accelerations

Has not tested what type of acceleration is going to be used.

OpenCL used to accelerate calculations

No OpenCL used/supported

Class to hold OpenCL devices

Device ID

Device type string

Device name string

Device vendor string

OpenCL version string

Execution capabilities of the device

Is device available?

Is device compiler available?

Maximum memory allocation size in bytes

Memory size in bytes

Maximum number of work-items that can be specified in each dimension of the work-group to clEnqueueNDRangeKernel.

Maximum number of work-items in a work-group executing a kernel using the data parallel execution model.

Constructor

Device ID

Class to hold OpenCL Platforms

Platform ID

OpenCL profile string. Profile name supported by the implementation.

OpenCL version string.

OpenCL name string.

OpenCL vendor string.

OpenCL extensions string.

Constructor.

Sets this platform's ID

Program related stuff

Local event

OpenCL context using all devices

Synchronous command queues that are executed in call order

Asynchronous command queues

Default synchronous command queue set as the first GPU, for ease of use.

Compiled program

Ends all commands being executed

Compiles program contained in a single string.

Source code to compile

Compiles program contained in a single string. Returns build logs for each device.

Source code to compile Build logs for each device

Compiles the program.

Source code to compile

Compiles the program. Returns the build logs for each device.

Source code array to compile Build logs for each device

Variables class

Data to be stored

Original variable length

Memory buffer

Constructor. Creates from an existing OpenCL variable

OpenCL variable pointer Original array length sizeOf(array datatype)

Constructor.

Variable whose size will be allocated in device memory.

Constructor.

Variable whose size will be allocated in device memory.

Constructor.

Variable whose size will be allocated in device memory.

Constructor.

Variable whose size will be allocated in device memory.

Constructor.

Variable whose size will be allocated in device memory.

Constructor.

Variable whose size will be allocated in device memory.

Writes variable to device

Values to write to device Command queue to use TRUE to return only after completed writing. OpenCL Event associated to this operation OpenCL Events that need to finish before this one can start

Writes variable to device

Values to write to device

Writes variable to device

Values to write to device Command queue to use TRUE to return only after completed writing. OpenCL Event associated to this operation OpenCL Events that need to finish before this one can start

Writes variable to device

Values to write to device

Writes variable to device

Values to write to device Command queue to use TRUE to return only after completed writing. OpenCL Event associated to this operation OpenCL Events that need to finish before this one can start

Writes variable to device

Values to write to device

Writes variable to device

Values to write to device Command queue to use TRUE to return only after completed writing. OpenCL Event associated to this operation OpenCL Events that need to finish before this one can start

Writes variable to device

Values to write to device

Writes variable to device

Values to write to device Command queue to use TRUE to return only after completed writing. OpenCL Event associated to this operation OpenCL Events that need to finish before this one can start

Writes variable to device

Values to write to device

Writes variable to device

Values to write to device Command queue to use TRUE to return only after completed writing. OpenCL Event associated to this operation OpenCL Events that need to finish before this one can start

Writes variable to device

Values to write to device

Reads variable from device.

Values to store data coming from device Command queue to use TRUE to return only after completed reading. OpenCL Event associated with this operation OpenCL Events that need to finish before this one can start

Reads variable from device. Does not return until data has been copied.

Values to store data coming from device

Reads variable from device.

Values to store data coming from device Command queue to use TRUE to return only after completed reading. OpenCL Event associated with this operation OpenCL Events that need to finish before this one can start

Reads variable from device. Does not return until data has been copied.

Values to store data coming from device

Reads variable from device.

Values to store data coming from device Command queue to use TRUE to return only after completed reading. OpenCL Event associated with this operation OpenCL Events that need to finish before this one can start

Reads variable from device. Does not return until data has been copied.

Values to store data coming from device

Reads variable from device.

Values to store data coming from device Command queue to use TRUE to return only after completed reading. OpenCL Event associated with this operation OpenCL Events that need to finish before this one can start

Reads variable from device. Does not return until data has been copied.

Values to store data coming from device

Reads variable from device.

Values to store data coming from device Command queue to use TRUE to return only after completed reading. OpenCL Event associated with this operation OpenCL Events that need to finish before this one can start

Reads variable from device. Does not return until data has been copied.

Values to store data coming from device

Reads variable from device.

Values to store data coming from device Command queue to use TRUE to return only after completed reading. OpenCL Event associated with this operation OpenCL Events that need to finish before this one can start

Reads variable from device. Does not return until data has been copied.

Values to store data coming from device

Sets this variable as an argument for a kernel

Index of kernel argument Kernel to receive argument

Releases variable from memory.

Destructor

Returns the size of the stored variable

Kernels class

Local kernel storage

Number of arguments

Creates a new Kernel

"Remember" variables

Sets kernel arguments

Variables to be set as arguments

Execute this kernel

Command queue to use Arguments of the kernel function Array of maximum index arrays. Total work-items = product(max[i],i+0..n-1), n=max.Length Events to wait before executing this Event of this command

Execute this kernel

Command queue to use Arguments of the kernel function Array of maximum index arrays. Total work-items = product(max[i],i+0..n-1), n=max.Length Local work sizes Events to wait before executing this Event of this command

Execute this kernel

Array of maximum index arrays. Total work-items = product(max[i],i+0..n-1), n=max.Length Arguments of the kernel function

Execute this kernel using work_dim = 1

Global work size in one-dimension. global_work_size = new int[1] {GlobalWorkSize} Arguments of the kernel function

Execute this kernel

Array of maximum index arrays. Total work-items = product(max[i],i+0..n-1), n=max.Length Local work sizes Arguments of the kernel function

Releases kernel from memory

Destructor

Gets how many arguments this kernel has

OpenCL programs

Basic linear algebra functions

Float vector sum kernel

float matrix multiplication kernel

float Gauss Seidel method

LU factorizaton method

Constructor. Builds OpenCL program.

Converts vector to matrix

Vector Matrix first dimension Matrix second dimension

Converts matrix to vector

Matrix Matrix first dimension Matrix second dimension

Returns the sum of two matrices

Matrix 1 Matrix 2

Matrix multiplication

Matrix 1 Matrix 2

Gauss Seidel method for iterative linear system solving. Returns unknown x

Matrix M so that Mx=b Initial estimate Known vector b Gauss-Seidel iterations per step Maximum number of times Gauss-Seidel iterations Desired sqrt(Sum(error[i]^2))*number of equations Estimated absolute error per component

Gauss Seidel method for iterative linear system solving. Returns unknown x

Matrix M so that Mx=b Initial estimate Known vector b Estimated error per equation

Solves linear system Mx = b by LU decomposition. Returns x

Matrix M Vector b Maximum acceptable absolute error Maximum iterations

Calculates LU decomposition of M matrix

Matrix to decompose Matrix dimension Swap index

Gauss Seidel method. Make sure to send x = b. Replaces x.

Matrix multiplication. Dimensions { p, r }.

Basic linear algebra functions

Float vector sum kernel

double matrix multiplication kernel

double Gauss Seidel method

LU factorizaton method

Constructor. Builds OpenCL program.

Converts vector to matrix

Vector Matrix first dimension Matrix second dimension

Converts matrix to vector

Matrix Matrix first dimension Matrix second dimension

Returns the sum of two matrices

Matrix 1 Matrix 2

Matrix multiplication

Matrix 1 Matrix 2

Gauss Seidel method for iterative linear system solving. Returns unknown x

Matrix M so that Mx=b Initial estimate Known vector b Gauss-Seidel iterations per step Maximum number of times Gauss-Seidel iterations Desired sqrt(Sum(error[i]^2))*number of equations Estimated absolute error per component

Gauss Seidel method for iterative linear system solving. Returns unknown x

Matrix M so that Mx=b Initial estimate Known vector b Estimated error per equation

Solves linear system Mx = b by LU decomposition. Returns x

Matrix M Vector b Maximum acceptable absolute error Maximum iterations

Calculates LU decomposition of M matrix

Matrix to decompose Matrix dimension Swap index

Gauss Seidel method. Make sure to send x = b. Replaces x.

Matrix multiplication. Dimensions { p, r }.

Float differential equation integrator

Writes final Y values and estimated absolute errors

Updates X to current time

Independent variable current value in OpenCL memory

Dynamic system current state in OpenCL memory

Derivative calculator

Constructor.

Initial state of system Desired step per integration pass Initial independent variable value Function to calculate derivatives vector

Integrates equation set to a final value using current stepsize. Ideally, final value and currentX should multiples of stepsize.

Final value to reach.

Sets current state

New independent variable value New state values

Takes an integration step. Saves and returns stepsize back to what it was.

Step size to use

Takes an integration step

Sets step size.

Step size to use

Gets current values of space-state variables (from Device).

Gets current absolute error sum

Gets current independent variable value (from Device).

Function to calculate derivatives vector

IN: Scalar. Independent variable. IN: State-space vector. OUT: Derivatives

OpenCL source

double differential equation integrator

Writes final Y values and estimated absolute errors

Updates X to current time

Independent variable current value in OpenCL memory

Dynamic system current state in OpenCL memory

Derivative calculator

Constructor.

Initial state of system Desired step per integration pass Initial independent variable value Function to calculate derivatives vector

Integrates equation set to a final value using current stepsize. Ideally, final value and currentX should multiples of stepsize.

Final value to reach.

Sets current state

New independent variable value New state values

Takes an integration step. Saves and returns stepsize back to what it was.

Step size to use

Takes an integration step

Sets step size.

Step size to use

Gets current values of space-state variables (from Device).

Gets current absolute error sum

Gets current independent variable value (from Device).

Function to calculate derivatives vector

IN: Scalar. Independent variable. IN: State-space vector. OUT: Derivatives

OpenCL source

Discrete element modeling. Calculates derivatives of n particle-spring model into a 6n space-state system (positions, velocities, x,y,z each).

Mass values (n)

Original positions (3n)

Origins (L) origs[i] connects to dests[i]

Destinations (L)

Spring constants (L)

Spring constants to ground (n)

Damping (L)

Damping to ground (n)

Number of Connections (1)

Initial distances (L)

Actuating forces (3*n)

Connection forces (L)

Nodes connections (int, 20*n)

Initial lengths kernel. work_dim = 1, globalsize = n

Initial length arguments

Reset forces kernel. work_dim = 1, globalsize = 3n

Reset forces arguments

Calculate forces kernel. work_dim = 1, globalsize = L

Calculate forces arguments

Calculate forces kernel. work_dim = 1, globalsize = n

Calculate forces arguments

Calculate forces kernel. work_dim = 1, globalsize = L

Calculate forces arguments

Calculate nodes connections. work_dim = 1, globalsize = n

Calculate nodes connections arguments

Constructor.

Number of masses in the system Number of connections Mass of each vertex Position and velocity of vertexes [2*3*i] - posx, [2*(3*i+1)] - posy, [2*(3*i+2)] - posz, [1+2*3*i] - velx, [1+2*(3*i+1)] - vely, [1+2*(3*i+2)] - velz Origin vertex of connections. Spring connects Origin[i] to Dests[i] Destination vertex of connections. Spring connects Origin[i] to Dests[i] Spring constant for each connection Spring constant for each mass, connecting to ground (nMass) Structural damping (relative-speed dependant) (nConnections) Absolute damping proportional to speed relative to Earth (nMass)

Calculates derivatives of deformable body space-state vector dydx[6n]. dydx[2i] - i-th position deriv, dydx[2i+1] - ith velocity deriv

Reset forces. Work_dim = 1, nmax = { nMasses }

Derivatives sketch. Work_dim = 1, nmax = { 3 * nMasses }

Forces Masses Independent variable State space vector Derivatives

Initial L0 calculation. work_dim = 1, global_work_size[0]=nConnections

Forces calculation. Returns forces. Work_dim = 1, nmax = { nConnections }

Calculates forces to ground. w_dim=1, global_work_size = nMasses

Calculates forces to ground. w_dim=1, global_work_size = nMasses

Vector sum - sums two vectors

Integer vector sum. work_dim = 1

Float vector sum. work_dim=1

Long vector sum. work_dim=1

Double vector sum. work_dim=1

Minimum differences - makes matrix of shifted differences of vectors

Differences convolution. work_dim=2

Floating point particle system physics

Motion Newton-law 1D solver. Kernel: rk46

Force applier to particles. Kernels: ResetForces, ApplyGravity, FloorCollision

Collision applier to particles. Kernels: ResetForces, ApplyGravity, FloorCollision

Initializes physics program. Components indexes: [i] - x, [i+1] - y, [i+2] - z

Number of particles

Sets particles parameters

Positions (3*numParticles) Speeds (3*numParticles) Masses (numParticles) Collision sizes (numParticles)

Gets particles positions

Gets how many close neighbors a particle has. Use this to avoid drawing unnecessary particles

Gets simulation time

Stepsize

Executes an integration step

Motion step arguments

Takes an integration step

Step size

Applies gravity

Apply gravity arguments

Applies gravity force.

Gravity force. Remember to use negative for down direction.

Clear forces

Apply gravity arguments

Clears forces

Floor collision

Apply floor collision arguments

Applies floor collision?

Wall collision

Apply floor collision arguments

Applies floor collision?

Self collision

Apply self collision arguments

Applies self collision?

Reset close neighbors

Apply self collision arguments

An useful class to read/write/delete/count registry keys

To read a registry key. input: KeyName (string) output: value (string)

To write into a registry key. input: KeyName (string) , Value (object) output: true or false

To delete a registry key. input: KeyName (string) output: true or false

To delete a sub key and any child. input: void output: true or false

Retrive the count of subkeys at the current key. input: void output: number of subkeys

Retrive the count of values in the key. input: void output: number of keys

A property to show or hide error messages (default = false)

A property to set the SubKey value (default = "SOFTWARE\\" + Application.ProductName.ToUpper())

A property to set the BaseRegistryKey value. (default = Registry.LocalMachine)

A strongly-typed resource class, for looking up localized strings, etc.

Returns the cached ResourceManager instance used by this class.

Overrides the current thread's CurrentUICulture property for all resource lookups using this strongly typed resource class.

Creates a Matrix of real numbers.

Matrix items

Linear Solve parameter. Maximum mean error allowable to linear solve method.

Linear Solve parameter. Limit linear solution correction iterations.

Linear Solve parameter. Should the solution method halt if a hard singulariry is found in matrix?

Linear Solve parameter. Should the method ignore if the matrix has a close-to-zero determinant and keep solving?

Constructor. Initializes a [0,0] matrix.

Constructor. Creates matrix from existing items.

Matrix items to create matrix from.

Copy constructor.

Matrix to copy from.

Constructor. Creates empty matrix with specified dimensions.

Number of rows in matrix. Number of columns in matrix.

Sums two matrixes.

First matrix to sum. Second matrix to sum.

Subtracts two matrixes.

Matrix to subtract from. Matrix to be subtracted.

Matrix multiplication. Notice that m1 rows should be the same as m2 lines for compatibility.

First matrix to multiply. Second matrix to multiply.

Matrix scalar multiplication.

Matrix to multiply. Scalar to multiply.

Matrix scalar multiplication.

Matrix to multiply. Scalar to multiply.

Matrix post-vector multiplication. Notice that a vector is a [1,Cols] matrix which means vector length should be equal matrix number of columns.

Matrix to multiply. vector to multiply.

Matrix pre-vector multiplication. Notice that a vector is a [1,Cols] matrix which means vector length should be equal matrix number of lines.

Matrix to multiply. vector to multiply.

Matrix scalar division.

Matrix to multiply. Scalar to divide each element of matrix.

Compares matrixes and returns true if they are identical.

Matrix to compare to.

Returns matrix transpose.

Element-wise product. This is not regular matrix product. It multiplies elements at corresponding positions.

Matrix to multiply element-wise.

Returns Euclidean norm of the matrix.

Dot product of two matrixes.

Matrix to dot product with/

Element-wise inversion. Returns the matrix with each element (x) inverted (1/x).

Creates internal LU factorization of this matrix.

Returns the solution x to the linear system Ax=b, A being this matrix.

Right-hand side known values.

Returns the determinant of this matrix.

Returns the inverse of this matrix.

Returns the solution x to the linear system A'Ax=A'b, A being this matrix.

Right-hand side known values.

Returns the weighted solution x to the linear system A'WAx=A'Wb, A being this matrix. TO DO: Correct quality check

Right-hand side known values. Weight matrix.

Returns the weighted solution x to the linear system A'DAx=A'Db, A being this matrix. D is a diagonal weight matrix.

Right-hand side known values. Main diagonal elements of diagonal weight matrix D.

Calculates R², corrected R² and Quadratic Error for the trySolution x to the linear system A'Ax=A'b, A being this matrix.

Right-hand side known values. Solution to use to evaluate quality indexers.

Applies the Gram-Schmidt orthonormalization method to this matrix, replacing it by the orthonormalized matrix.

Applies the Gram-Schmidt orthonormalization method to this matrix, replacing it by the orthonormalized matrix and also correcting right-hand Y values for a linear system solve.

Right-hand side known values.

Applies the Gram-Schmidt orthonormalization method to this matrix using a pre-set order of normalization. Replaces current matrix by the orthonormalized matrix and also correcting right-hand Y values for a linear system solve.

Right-hand side known values. Line order to apply the orthonormalization method. Starts orthonormalization from line lineOrder[indStart]. Assumes previous lines are already normalized.

Returns a string representing this matrix.

Accesses items in current matrix.

Row of element to access. Column of element to access.

Gets the number of rows in this matrix.

Gets the number of columns in this matrix.

Returns the R² index of last fit.

Returns corrected R² index of last fit.

Returns the sum of quadratic errors of last fit.