Accessing new heap costs more than old?
ahey at iee.org
Mon Sep 5 03:11:37 EDT 2005
On Saturday 03 Sep 2005 6:14 pm, Adrian Hey wrote:
> But I can't think of a plausible explanation for this. The overall heap
> burn rate should be about the same in each case, as should the overall
> amount of live (non-garbage) heap records.
Hmm.. A little more thought leads me to suspect that this is an artifact
of generational garbage collection, which I guess is much faster if
the youngest generation is entirely garbage. Is that reasonable?
More information about the Glasgow-haskell-users