Comments:	Authenticated sender is <mert0407 AT sable DOT ox DOT ac DOT uk>
From:	"George Foot" <george DOT foot AT merton DOT oxford DOT ac DOT uk>
To:	DougEleveld <deleveld AT dds DOT nl>
Date:	Fri, 28 Aug 1998 19:03:25 +0000
MIME-Version:	1.0
Subject:	Re: question to malloc and free
Reply-to:	george DOT foot AT merton DOT oxford DOT ac DOT uk
CC:	djgpp AT delorie DOT com
Message-Id:	<E0zCStN-0005cG-00@sable.ox.ac.uk>

On 28 Aug 98 at 18:03, DougEleveld wrote:

> Endlisnis wrote:
> 
> >     I've read that the size of the block allocated by malloc/new is always
> > a power of 2, and rounded up to one if it isn't (after adding it's few
> > words).  Why does it do this?  This seems (to me) like it would waste about
> > 1/2 the RAM allocated.
> 
> This way the memory actually requested is of a known size and the
> allocations can be done _very_ fast.  It's the clasic size/speed
> tradeoff.  But remember (I think) that only addressing space from the
> 4Gig is wasted, not actual memory.  (!Simplified explanation!) Because
> if you allocated 1.2k and get 2k, and you never access out of the range
> that you are allowed (which you should never do anyway) then the other
> 0.8k will never get paged in, and it will make very little difference if
> 0.8k out of the 4gig is wasted, especially since maximum real+virtual
> memory is way less than 4gig anyway.

I think in fact the paging is done with 4k granularity, so a whole 4k
page is either in or out of physical memory at one time.  So if you
allocated 16k and only used less than 8k, the first two 4k pages
would be paged in and the remaining 8k would stay paged out.

IIRC the old malloc rounded up to 4k boundaries, so there was
significant wastage here if you allocated a lot of very small
structures (because the whole 4k page is paged in, and you only touch
a bit of it).  I might be misremembering though, I don't have the
sources to that malloc.

-- 
george DOT foot AT merton DOT oxford DOT ac DOT uk

- Raw text -