RAID 0 doesn't add any extra processing burden on the CPU other than the few instructions required to decide which drive the data is located on. Unlike redundant RAID levels (such as 1, 0+1, 5, etc), RAID 0 doesn't require any extra I/Os than for non-RAID drives, and there are no parity calculations either.
Compared to the instruction burden required by the file system just to figure out where the appropriate data is, and the work done by the device driver to translate file system calls into actual I/O requests for the drive, those extra couple of instructions needed to decide which drive to access are completely insignificant.
For RAID 0, there's no appreciable overhead in using the motherboard chipset RAID solution.
Actually, in modern processors even RAID 5 (which requires a lot more work than RAID 0 to run in software) only consumes a few percent of CPU time.