brian2cuda package

Package implementing the CUDA “standalone” Device and CodeObject.

Functions

example_run([device_name, directory])

Run a simple example simulation to test whether Brian2CUDA is correctly set up.

_version module

Functions

get_versions()

binomial module

CUDA implementation of BinomialFunction

codeobject module

Module implementing the CUDA “standalone” CodeObject. Brian2CUDA implements two different code objects. CUDAStandaloneCodeObject is the standard implementation, which does not use atomic operations but serialized synaptic effect application if race conditions are possible. CUDAStandaloneAtomicsCodeObject uses atomic operations which allows parallel effect applications even when race conditions are possible.

Exported members: CUDAStandaloneCodeObject, CUDAStandaloneAtomicsCodeObject

Classes

CUDAStandaloneAtomicsCodeObject(*args, **kw)

CUDA standalone code object which uses atomic operations for parallel execution

CUDAStandaloneCodeObject(*args, **kw)

CUDA standalone code object

cuda_generator module

Exported members: CUDACodeGenerator, CUDAAtomicsCodeGenerator, c_data_type

Classes

CUDAAtomicsCodeGenerator(*args, **kwds)

CUDACodeGenerator(*args, **kwds)

C++ language with CUDA library

ParallelisationError

cuda_prefs module

Preferences that relate to the brian2cuda interface.

Functions

validate_bundle_size_expression(string)

device module

Module implementing the CUDA “standalone” device.

Classes

CUDAStandaloneDevice()

The Device used for CUDA standalone simulations.

CUDAWriter(project_dir)

Methods

Functions

prepare_codeobj_code_for_rng(codeobj)

Prepare a CodeObject for random number generation (RNG).

Objects

cuda_standalone_device

The Device used for CUDA standalone simulations.

timedarray module

CUDA implementation of TimedArray

Subpackages