Skip to content

RichDet MaPMT class clean up

Christopher Rob Jones requested to merge RichDet-MaPMTCleanup into master

@seaso FYI. I will appreciate your review and comments.

This MR is primarily to start an overdue clean up of the MaPMT classes in RichDet, removing redundant code implementing features once considered for the upgrade but now no longer required. We need to clean up the code to streamline it, both in terms of runtime performance but also memory consumption (remember, we create quite a number of DeRichPMT PD instances so need to worry a bit about both how long it takes to initialize() them, and what there memory footprint is).

This MR starts this process. I have commented out (for now, will remove completely at some point) the code I know is not used. There is still more to do (for instance @seaso I assume I am correct that for Rich2 we only need to support the 'mixed' PMT option ?).

@seaso I believe you have also to some degree been looking into this at the same time as we. I suspect though your branch is not quite out of data with some recent updates I made to master, so probably the best approach is to collect all the cleanups in this branch.

Finally, there is also a small improvement to SIMD DeRichPMT::detectionPoint method, adding a native SIMD version (previously it just deferred to a loop calling the scalar version).

Edited by Christopher Rob Jones

Merge request reports