tree_d_3d.kernels.tests.test_kernels_emulated

The CUDA versions of the Barnes-Hut and FMM kernels against their Python versions.

Each CUDA kernel is compiled as C++ with the emulation stubs of test_cuda_kernels_emulated and run thread by thread, so the kernels are checked on machines without a GPU (the threads run one after another: atomic adds are exact sums, in another order than on the CPU).

Attributes

Name	Description
ORDER	No description.
ROOT	No description.
SOFTENING	No description.
THETA	No description.
TREE	No description.
pytestmark	No description.

Functions

Name	Description
log_tree	No description.
system	No description.
test_coincident_charges_are_reported	No description.
test_direct_sums	No description.
test_fmm_pipeline	Traversal, M2L, downward pass and evaluation, each against its Python version.
test_log_moments_and_evaluation	No description.
test_pair_energy	No description.
test_quadtree_build	The level-by-level build with the CUDA quadrant and scatter kernels.
test_quadtree_moments_and_evaluation	No description.
test_soften_pairs	No description.

ORDERattributemodule attribute#

View source

ORDER = 6

ROOTattributemodule attribute#

View source

ROOT = Quad(0.0, 0.0, 1.0 + 1e-09)

SOFTENINGattributemodule attribute#

View source

SOFTENING = 0.01

THETAattributemodule attribute#

View source

THETA = 0.6

TREEattributemodule attribute#

View source

TREE = {'leaf_capacity': 4, 'max_depth': 20}

pytestmarkattributemodule attribute#

View source

pytestmark = pytest.mark.skipif(COMPILER is None, reason='no C++ compiler')

log_treefunction#

View source

def log_tree(system)

systemfunction#

View source

def system()

test_coincident_charges_are_reportedfunction#

View source

def test_coincident_charges_are_reported(tmp_path, system)

test_direct_sumsfunction#

View source

def test_direct_sums(tmp_path, system, name)

test_fmm_pipelinefunction#

View source

def test_fmm_pipeline(tmp_path, system, log_tree)

Traversal, M2L, downward pass and evaluation, each against its Python version.

test_log_moments_and_evaluationfunction#

View source

def test_log_moments_and_evaluation(tmp_path, system, log_tree)

test_pair_energyfunction#

View source

def test_pair_energy(tmp_path, system)

test_quadtree_buildfunction#

View source

def test_quadtree_build(monkeypatch, system)

The level-by-level build with the CUDA quadrant and scatter kernels.

test_quadtree_moments_and_evaluationfunction#

View source

def test_quadtree_moments_and_evaluation(tmp_path, system)

test_soften_pairsfunction#

View source

def test_soften_pairs(tmp_path, system)