xphm is the first try, results are not reliable
xphm-adv is reliable, with srcmass1/2 prior
massl20 and three individual GW folders, 2 BBH + GW190426 are the final products. With rwalk2 dynesty and srcmchirp and q prior
===
190426 is workflow with rwalk2, not finish yet
dydefault, dydefault-128cpu is the same, default setting, different cpus
default-v2 increase nlive to 3000 and decrease dlogz to 0.05, works better
===
0430results stores the o3a results. Orginial inference outputs are from massl20, 190426-default-v2, two GW....BBH folders.