Date:	Sat, 17 Jul 1993 07:50:18 -0600
From:	jweiss AT silver DOT sdsmt DOT edu (John M. Weiss)
To:	dave AT echologic DOT com
Subject:	memory
Cc:	djgpp AT sun DOT soe DOT clarkson DOT edu

On 07/16/93 dave AT echologic DOT com (Dave Hayden) wrote:

> I saw your response to my idea of allocating a 2D array by chunks so
> as to minimizes paging.  As you point out, this alogrithm still
> depends on the amount of memory available, but I think you're wrong about
> it depending on the malloc algorithm.  I envision allocating all the
> space for the array in one malloc() call and just changing the way the
> memory is accessed.  If the memory is allocated in one call, the
> malloc algorithm is irrelevant.

You may be correct. I haven't thought through all the ramifications 
of this approach yet. It may be completely independent of malloc().

> As I write this, I thought of another tac you could try.  After
> allocating the array, scan it once by rows and once by columns, and
> time how long it takes.  Then do the same thing with a small portion
> of the array (maybe a 100x100 section).  If the times scale roughly,
> then you've probably got the array in memory...

I've thought about timing approaches too. There are a number 
of hacks I could use to determine when the system starts 
paging. I just wish there was a more straightforward method.

> Thanks for your time.  I may toy around with these ideas a little.
> I'll let you know what I find if you're interested.

Thanks for *your* time. I appreciate all the new ideas!

                                - JW

- Raw text -