search

petrelharp / ftprime_ms

4 stars 2 forks source link

possibly inaccurate y-axes in RAM usage figure #42

Closed ashander closed 6 years ago

ashander commented 6 years ago

@molpopgen are the y-axes in the right and left panes of the figure comparing RAM use without selection accurate?

(this is fig B2 in #41 but elsewhere in master; reproduced below.)

image

molpopgen commented 6 years ago

They're from the same files used for the timings. What seems amiss, specifically?

molpopgen commented 6 years ago

I can double check the code tomorrow I any case....

ashander commented 6 years ago

Ah, I was thrown by N=1e4 not being included in the simuPOP comparison.

So I think we need to make the point more clearly (or I need to remind myself) that even though our peak RAM use with pedigree tracking in simuPOP is insanely high, it's lower than without tracking.

image