Closed JLBLine closed 2 months ago
With 25cf255396b37c0f81a6f1afa008bafbb924c522, the function source_component_common::get_beam_gains_two_antennas
and is tested as working via cmake_testing/CUDA_code/source_components/test_source_component_common_two_antennas.c
. The test is working. Next, hook it into update_sum_visis_stokesI
.
In 53df7c997d7abe94e4b6af0dfc2d47f2f09e3c70 added cmake_testing/CUDA_code/source_components/test_update_sum_visis_multiants.c
which demonstrates that update_sum_visis_stokesI(QUV)
can now call get_beam_gains_two_antennas
. Have updated existing tests to still use one primary beam for all antennas using the same functions.
This will be a function that given a specific baseline, component, time, and frequency, chose the two primary beam values that need to be applied to the visibility. This makes sense to be a new function, as it will need to calculate two different indexes in the beam gain arrays instead of one (in the case of everything with the same primary beam).
Testing: Give the function an array of beam gain values (similar to
test_get_beam_gains.c
), but this time with say three antennas. Will need to supplyd_ant1_to_baseline_map
andd_ant2_to_baseline_map
on the CUDA side of things, so we can setup code to produce those arrays in the test. That can be copied intocalculate_visibilities::calculate_visibilities
later.