Rerun with the fortran strategies

[drvinceknight]

No description provided.

[marcharper]

I did do this a while back actually, generated all the match data, but did not upload new images. I can do it but there are a few new strategies now so a rerun is in order.

On my local machine the entire process takes multiple days with the extra strategies (it's basically quadratic in the number of strategies). However the match generation hasn't been parallelized: it's just two processes, one for the noisy matches and one for the regular ones. So it's probably a good time to fix that, and refactor to use the library's fingerprint methods. That should substantially improve the run time and would allow us to take advantage of the 16 core machines.

[drvinceknight]

Cool 👍 I must have forgotten you already did it :)

Refactoring sounds good. 👍

[drvinceknight]

FYI, I'm doing the fingerprints (both default and versus short run time) on the fingerprints repo but again that's (literally) taking weeks :)

[marcharper]

Well I never finished the generation of the figures so I didn't upload anything. I was intending to do the described refactor but it hasn't happened yet. I was also only using the short run time strategies.