> On a 386+ you can do REP MOVSD, which, given the same amount of
 > memory to move, would work twice as fast (on all processors, the
 > time it takes is determined by the count ([E]CX), not the size
 > being moved... moving 1,024 bytes with REP MOVSB takes the same
 > amount of time as moving 1,024 words with REP MOVSW.
Not exactly true.
When you are moving to an ISA device, such as most VGA's on 386 systems, the
data is being sent to the card a WORD at a time (ISA is 16-bit max)
 (barely any 386's have EISA, and none have PCI or VLB)
Also, if your REP MOVSD is misaligned, you again pick up problems even if you 
are writing to system memory, which then must break up your dwords into words 
(or even bytes) before writing.
--- GEcho 1.02+
---------------