You'll have to transfer the data with the CPU thus reducing availlable MIPS. DMA is done without CPU intervention.
Look at these examples (the third and the second) to see what you can do with DMA, and why you need it (as comparision, there are AVR logic analyzer projects out there, which reach up to a whopping 7kHz): http://www.cypress.com/?id=4391
danaaknight posted on 10 Sep 2012 12:31 PM PST
499 Forum Posts
DMA is a method, in today’s parts, using separate busses so that CPU intervention is
not required to transfer data from one peripheral to another, or I/O ports, etc.. It is
almost always faster than CPU directed transfers, so is capable of keeping up with
communication rates like SPI, UART...... It effectively gives you many more MIPs
in a machine, and depending on architecture runs concurrently with other CPU
As the name suggests, Direct Memory Access (DMA) block can be used to access the memory. This frees the CPU from the petty task of memory transfer and thereby focus on more critical functions to be performed.
You may refer to the application note http://www.cypress.com/?rID=37793 to know how to use this block.