I'm trying to use the _mem4 intrinsic to copy 4 bytes from one part of L2 SRAM to another part on the C674x.
Sadly, it only seems to work properly when both source and destination are 4-byte aligned. This is despite this intrinsic claiming to be *specifically* for unaligned access. Is this a bug with the SRAM and C674x?
When I exchange _mem4 for _amem4 my code behaves exactly the same which leads me to think that when SRAM is involved, there is no difference in these intrinsics.
Thanks.
