Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uwm.edu!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!mephisto!udel!new
From: new@udel.edu (Darren New)
Newsgroups: comp.sys.amiga.tech
Subject: Re: Accelerator and Maximum Fast RAM Query
Message-ID: <11902@nigel.udel.EDU>
Date: 22 Feb 90 22:04:54 GMT
References: <1179@mindlink.UUCP> <9786@cbmvax.commodore.com>
Sender: usenet@udel.EDU
Reply-To: new@udel.edu (Darren New)
Organization: University of Delaware
Lines: 12

In article <9786@cbmvax.commodore.com> peter@cbmvax.cbm.commodore.com (Peter Cherna) writes:
>Note that the system is set up so that memory allocation uses 32-bit memory
>before it moves on to the slower 16-bit memory, and that's why is doesn't
>"hurt" to have 16-bit memory around.

But note also:
If you have 16-bit memory then Avail() calls will access the 16-bit
memory.  I was doing an AvailMem() on every memory allocation
and it was almost as slow on a 68020 as it was on a 68000.  Eventually,
I just modified it to allocate a chunk of memory, then allocate
until it ran out, then free the chunk of memory.   I suggest
that if you are programming on a 68000, you do this too. -- Darren