Initial electr and next steps about egsnrc

[SimonBiggs]

I'm just on leave at the moment (only just got back into intermittent phone reception 🙂) but this looks perfect for me to make a priority to dig in and gets my hands dirty with when I get back. I'll be back early March ☺️.

[darcymason]

Sounds good, @SimonBiggs ... meanwhile I'm been plugging away at various subroutines, etc - annih, bhabha, brems... not hard to convert because mostly just math. After that, probably another round of tests to make sure they are all working correctly under various user settings, etc. I've started to realize that everything is so interconnected (heavily reliant on globals) that we probably can't do much on vectorizing or any other major code reorg without having almost everything converted over. I'm back to the simple target of a 'first pass' conversion where everything works on Python, but in pretty much identical fashion to the original, ... then we can be in a good position for restructuring.

[SimonBiggs]

meanwhile I'm been plugging away at various subroutines

If you find yourself looking for what to prioritise next, the section of egs that does some of the following stages might be worth a look in:

SimonBiggs/pyegsnrc#1

But only if there isn't something else that is a natural next step.

we probably can't do much on vectorizing or any other major code reorg without having almost everything converted over

Kk

[darcymason]

If you find yourself looking for what to prioritise next, the section of egs that does some of the following stages might be worth a look in:

Sure, I'll have a look and see what I can pick up - Maybe I can look at the main photon loop and try to summarize the random sampling towards the different interactions.

[darcymason]

Closing as this is old information -- no longer using call-outs to Fortran code with the current code and GPU use.