>>>>> "Anthony" == Anthony Preston <apreston@isd.csc.com> writes:

>>  Hey, there are several non-contiguous memory patches out there, if
>> you can compile your own kernel.  Patches have floated on amiga-dev
>> at several times.  Either go to an archive (preferred) or ask me
>> for a multisegmented memory patch.  Those two extra megs *will*
>> make a difference as they will be used for the kernel which is the
>> most important code to execute fast.  Just remember to use -p as
>> option to loadbsd, or otherwise the kernel will end up in the
>> 16-bit memory.
>> 
>> Niklas
>> 
Anthony>   Wait a minute... The VXL-30 memory autoconfigures as the
Anthony> first 2 MBs of fast memory, then the rest of the 6 MBs
Anthony> autofigures for 8 MBs of fast.  The memory is contiguous, I
Anthony> don't have the VXL-30 memory mapped high(although if I could
Anthony> get the 1 MB x 4 chips, I would have 8 MB 32 bit mapped high
Anthony> and 8 MB 16 bit fast).

Well, there are going to be two memory chunks anyway, with differet
priorities.  However, with this setup you have yet another way to go
(of course the multiple chunk patches still work), you can use
mergemem from 1.3.  It'll merge your two contiguous chunks so they'll
look as one.

Niklas