eprdata14 XS Data: Faulty? #81

Update: These two problems are without a doubt coupled:
https://github.com/openmc-dev/data/issues/81
https://github.com/openmc-dev/openmc/issues/2421

It is likely the division errors in the from_ace routine

I believe there is an issue with the data converted from the LANL MCNP eprdata14 ACE file.
This affects ENDF7.x LANL libraries
https://openmc.org/lanl-data-libraries/

It appears that several of the eprdata14 XS are not being 'leveraged' in OpenMC; in my problem a number of the photons undergo far fewer reactions than with other libraries.

When I convert EPICS2017 myself, EPICS2014 myself and use them with ENDF80 neutron data I get the same answer as with Official release OpenMC ENDF80 np. Which is significantly different than with eprdata14 (using OpenMC official endf80 neutron data).
I have not been able to convert eprdata14 ACE myself, see openmc-dev/openmc#2421

Running the same problem in MCNP6.2 with mcplib84, eprdata14 and jendl50 (EPICS2017) I get the same results in all. No major deviation for eprdata14

I am running OpenMC 0.13.3-dev.

Here are some results with the photon_source test.
Slightly modified (vaccuum conditions, 10 batches, Absorption score), see attached model.zip