cudolfinx.assemble

GPU-accelerated assembly routines.

Functions

create_petsc_cuda_vector(L)

Create PETSc Vector on device.

Classes

CUDAAssembler()

Class for assembly on the GPU.
class cudolfinx.assemble.CUDAAssembler[source]

Bases: object

Class for assembly on the GPU.

Initialize the assembler.

apply_lifting(b: CUDAVector, a: list[CUDAForm], bcs: list[list[DirichletBC] | CUDADirichletBC], x0: list[CUDAVector] | None = None, scale: float = 1.0, coeffs: list[list[Function]] | None = None)[source]

GPU equivalent of apply_lifting.

Parameters:

  • b – CUDAVector to modify

  • a – list of forms to lift

  • bcs – list of boundary condition lists

  • x0 – optional list of shift vectors for lifting

  • scale – scale of lifting

  • coeffs – coefficients to (re-)pack

apply_lifting_block(b: CUDAVector, a: BlockCUDAForm, bcs: list[DirichletBC] | Any, x0: CUDAVector | None = None, alpha: float = 1.0, set_bcs=True)[source]

Apply lifting with a single block form.

assemble_matrix(a: CUDAForm, mat: CUDAMatrix | None = None, bcs: list[DirichletBC] | CUDADirichletBC | None = [], diagonal: float = 1.0, constants: list | None = None, coeffs: list | None = None, zero: bool = True)[source]

Assemble bilinear form into a matrix on the GPU.

Parameters:

  • a – The bilinear form to assemble.

  • mat – Matrix in which to store assembled values.

  • bcs – Boundary conditions that affect the assembled matrix. Degrees-of-freedom constrained by a boundary condition will have their rows/columns zeroed and the value diagonal set on on the matrix diagonal.

  • diagonal – Value to set on the diagonal.

  • constants – Constants that appear in the form. If not provided, any required coefficients will be computed.

  • coeffs – Optional list of form coefficients to repack. If not specified, all will be repacked. If an empty list is passed, no repacking will be performed.

  • zero – Whether to zero the matrix prior to assembly

Returns:

Matrix representation of the bilinear form a.

Note

The returned matrix is not finalised, i.e. ghost values are not accumulated.

assemble_matrix_block(a: BlockCUDAForm, mat: CUDAMatrix | None = None, bcs: list[Any] | None = None, coeffs: list[list[Any]] | None = None, constants: list[list[Any]] | None = None, zero: bool = True)[source]

Assemble block form into a matrix on the GPU.

assemble_scalar(b: CUDAForm, constants=None, coeffs=None)[source]

Assemble scalar integral on GPU.

Parameters:

  • b – the functional to use for assembly

  • constants – Form constants

  • coeffs – Optional list of form coefficients to repack. If not specified, all will be repacked. If an empty list is passed, no repacking will be performed.

assemble_vector(b: CUDAForm, vec: CUDAVector | None = None, constants=None, coeffs=None, zero=True)[source]

Assemble linear form into vector on GPU.

Parameters:

  • b – the linear form to use for assembly

  • vec – the vector to assemble into. Created if it doesn’t exist

  • constants – Form constants

  • coeffs – Optional list of form coefficients to repack. If not specified, all will be repacked. If an empty list is passed, no repacking will be performed.

  • zero – Whether to zero vector entries before assembly (defaults to true)

assemble_vector_block(b: BlockCUDAForm, vec: CUDAVector | None = None, constants=None, coeffs=None, zero=True)[source]

Assemble block linear form into vector on GPU.

Parameters:

  • b – the block linear form to use for assembly

  • vec – the vector to assemble into. Created if not provided

  • constants – List of constants for each sub form

  • coeffs –

    Optional list of form coefficients to repack.

    If not specified, all will be repacked.

    If an empty list is passed, no repacking will be performed.

  • zero – Whether to zero vector entries before assembly (defaults to true)

create_matrix(a: CUDAForm) → CUDAMatrix[source]

Create a CUDAMatrix from a given form.

create_matrix_block(a: BlockCUDAForm) → CUDAMatrix[source]

Create a block matrix from a block form.

create_vector(b: CUDAForm) → CUDAVector[source]

Create a CUDAVector from a given form.

create_vector_block(b: BlockCUDAForm) → CUDAVector[source]

Create a CUDAVector from a given form.

pack_bcs(bcs: list[DirichletBC]) → CUDADirichletBC[source]

Pack boundary conditions into a single object for use in assembly.

The returned object is of type CUDADirichletBC and can be used in place of a list of regular DirichletBCs. This is more efficient when performing multiple operations with the same list of boundary conditions, or when boundary condition values need to change over time.

pack_coefficients(a: CUDAForm, coefficients: list[Function] | None = None)[source]

Pack coefficients on device.

set_bc(b: CUDAVector, bcs: list[DirichletBC] | CUDADirichletBC, V: FunctionSpace, x0: CUDAVector | None = None, scale: float = 1.0)[source]

Set boundary conditions on device.

Parameters:

  • b – vector to modify

  • bcs – collection of bcs

  • V – FunctionSpace to which bcs apply

  • x0 – optional shift vector

  • scale – scaling factor

cudolfinx.assemble.create_petsc_cuda_vector(L: Form) → Vec[source]

Create PETSc Vector on device.