Erwan Pannier

Indeed I could reproduce it too. @CorentinGrimaldi may be due to the Slit errors when using a dispersion, that you spotted already. @TranHuuNhatHuy let's ignore for the moment : the...

Hello, thanks for contributing. For some reason the test suite wasn't started (maybe because yo'ure a first time contributor). I'll start it manually by overwritting one of your changes

It is supposed to happen automatically, nothing needed on your side. Happened to us recently (https://github.com/radis/radis/pull/426#issuecomment-1068317814) . @gagan-aryan I think you fixed it last time. Can you try to push...

Error in tests https://app.travis-ci.com/github/radis/radis/jobs/571013983#L1815 seen to be due to the different value of get_power() Given that the new one is more physically accurate we should update the test reference value

Btw @AllanHOlesenBW , we'll release a new Radis version soon. To implement your changes, can you update the hardcoded power value to the new (corrected) value ?

Yes, it is this value in `radis/test/spectrum/test_spectrum.py` : https://app.travis-ci.com/github/radis/radis/jobs/571013983#L1815 (the value is expected to change since the old method was using the wrong wavelengths)

@AllanHOlesenBW I'm delaying this to the next release (0.14), please tell us if you need help updating the hardcoded value to pass the tests

@AllanHOlesenBW if you need us to update the value, please tell us !

Hello @dcmvdbekerom , thanks for writing this exhaustive post !! @vmol on Slack may be able to help on the "Eliminate fp32 atomic-adds" ? I'm surprised that only CO2 is...

One thing I'm not sure so sure is if the current GPU draft implementation has both resonant [self] and external [usually : air] pressure broadening ?