[MLton-devel] mark-compact-ratio

[Stephen Weeks]

> I'm surprised by Steve's results on mcr ratio.  What is the intuition behind
> these numbers?

My explanation is that paging really hurts, so it's good to avoid it
as much as possible.  Having a lower mcr decreases paging costs by
using less memory.

> I'm also puzzled by why gc times drop so significantly as
> the mcr ratio goes up; isn't it the case that gc cost should be proportional
> (not inversely proportional) to total heap size?

No.  Having a lower mcr means that you have to do gcs more frequently,
and hence spend more time in gc.  Fortunately, a lot of these are
minor gcs and so the cost doesn't go through the roof.  We would have
seen different results with just the mark-compact gc but no
generational gc.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone?  Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel