[kaffe] garbage collection strategy

[yan wei]

i read the article about kaffe's FAQ.gcstrategy, a paragraph is like this:
.......
So, what does Kaffe do to find the sweet spot between not collecting too
often and not using too much memory?  It looks at how much memory has been
allocated since the last time a collection happened.  If this amount of
memory is less than 1/3 of the total amount of memory in use, then the
collection is skipped.  In our hypothetical example above, Kaffe would
use 24MB (since it will grow to 24MB in 1MB increments.)  Every time 8MB
of short-lived data have been added to the long-lived data occupying 16MB,
kaffe will collect and free 8MB.  Hence, it will perform 160/(24-16) = 20
collections instead of 160 or 4.
....

my question is: how to get the 24MB? 64/3=22, it should be 23MB since it
increments 1MB I think.
can anybody help me?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://kaffe.org/pipermail/kaffe/attachments/20060302/bcef3d60/attachment-0005.htm