[MLton-devel] Stop-and-copy / mark-compact switch criterion

Stephen Weeks MLton@mlton.org
Fri, 16 Aug 2002 16:38:46 -0700


> Not  that  I know either way, but I think that making the nursery that big is
> probably sub-optimal.  If you have a lot of RAM, then the nursery is going to
> be  large  enough that it is constantly going to be fetched from main memory.
> Also, since you walk though it, you are going to push other things out.   The
> main argument in the other direction is that the nursery access is sequential
> for allocation so the misses aren't as expensive.

I figure the sequential access making misses cheaper, plus all the
other benefits I mentioned in the last email make a large nursery a
win.

One other point to keep in mind (that you couldn't possibly have
known until I mention it) is that I only plan on using generational gc
when the live ratio is small (say < 4), meaning that we're already in
a mode where there's a lot of live data and we weren't able to get the
desired live ratio (8).


-------------------------------------------------------
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