Slightly reorder operations in calculation of apar_m to improve cancellation
David Dickinson
Branch: micro/small_change_to_apar_m_operation_order_to_improve_cancellation
Branch: 8.0.4-RC
Merged
Merged pull request
Merged in micro/small_change_to_apar_m_operation_order_to_improve_cancellation (pull request #280)
For half the calls to set_source we'd expect apar = aparnew and hence apar_m (= dA/dt) == 0. However, with the existing operation order this cancellation would not occur exactly. This is unlikely to have a significant impact on anything but can help with some detailed testing.