FIXME : Add documentation
| Type | Intent | Optional | Attributes | Name | ||
|---|---|---|---|---|---|---|
| complex, | intent(in), | dimension (-ntgrid:,:,g_lo%llim_proc:) | :: | g | ||
| complex, | intent(out), | dimension (-ntgrid:,:,:,:) | :: | total | ||
| logical, | intent(in), | optional | :: | all | ||
| logical, | intent(in), | optional | :: | full_arr |
Takes f = f(theta, sigma ; x, y, z, lambda, E, species) and returns int f, where the integral is over all velocity space returns results to PE 0 [or to all processors if 'all' is present in input arg list]
| Type | Intent | Optional | Attributes | Name | ||
|---|---|---|---|---|---|---|
| real, | intent(in), | dimension (-ntgrid:,:,g_lo%llim_proc:) | :: | g | ||
| real, | intent(out), | dimension (-ntgrid:,:,:,:) | :: | total | ||
| logical, | intent(in), | optional | :: | all | ||
| logical, | intent(in), | optional | :: | full_arr |
Perform an integral over velocity space whilst in the LE_LAYOUT in which we have ensured that all of velocity space is local. As such we don't need any calls to MPI reduction routines. Note that this means the processors for different distributed spatial points (x,y) don't know the results at other points.
| Type | Intent | Optional | Attributes | Name | ||
|---|---|---|---|---|---|---|
| type(le_layout_type), | intent(in) | :: | lo | |||
| complex, | intent(in), | dimension (:,:,lo%llim_proc:) | :: | g | ||
| complex, | intent(out), | dimension (lo%llim_proc:) | :: | total |
FIXME : Add documentation
| Type | Intent | Optional | Attributes | Name | ||
|---|---|---|---|---|---|---|
| real, | intent(in), | dimension (-ntgrid:,:,p_lo%llim_proc:) | :: | g | ||
| real, | intent(out), | dimension (-ntgrid:,:,:) | :: | total | ||
| integer, | intent(in), | optional | :: | all |