General Question: 

Is there a mechanism in the kernel by which particular chunks of physical 
memory are given priority for use over others based on some criteria? 

Specific Question: 

On certain architectures (VME Bus systems in general, The Heurikon VME532 
and MVME172 in particular), there is relatively fast local memory, and 
relatively slow expansion memory. 

For example, on the VME532 (NS32532 @ 25MHz), there is 4MB of zero-wait 
local memory.  On the MVME172 (MC68060 @ 50MHz) that I have, there is 32MB 
of (I think) 1-wait local memory.  I can add an arbitratry amount of VME 
memory, but it's significantly slower than the local memory.  BTW...neither 
of these is VSB bus memory.  That would be faster, but I don't know how much 
faster. 

Is there a way, implimented or theoretical, that would maximize the 
performance of systems with different physical memory performance 
characteristics? 

Ken