People often say that focusing on one (CPU or GPU) project is most profitable. According to how I understand the calculation of MAG, this should not be the case. The magnitude for a project is calculated

In the search of the most mag/credit-efficient BOINC project for your computer hardware, many people select the projects using for example these nice tables from @nexusprime: You can take this even further,

There was recently a discussion on Reddit ( about the scientific aspects of Moo Wrapper, which is the BOINC-version of the RC5-72 challenge. This discussion led to the current poll on the question, whether