On my MacBook I get slightly different results from the astrolib celestial benchmarks just by re-running them (i.e. git diff shows all output tables as modified, which doesn't happen for any other tool).
This has already been observed and discussed here:
#32 (comment)
It's pretty annoying and I don't want to spend time debugging this astrolib issue.
My suggestion would be to disable astrolib for now from the coordinates-benchmark and to only investigate this issue if we really need it, i.e. if it offers something the other packages we compare against don't.
@astrofrog What do you think?
On my MacBook I get slightly different results from the
astrolibcelestial benchmarks just by re-running them (i.e.git diffshows all output tables as modified, which doesn't happen for any other tool).This has already been observed and discussed here:
#32 (comment)
It's pretty annoying and I don't want to spend time debugging this
astrolibissue.My suggestion would be to disable
astrolibfor now from thecoordinates-benchmarkand to only investigate this issue if we really need it, i.e. if it offers something the other packages we compare against don't.@astrofrog What do you think?