What is difference between mem copy and full copy in lmbench3?

Question

I am trying to understand the bandwidth benchmarking. I came across these routines MCP and FCP in lmbench3. I disassembled the binary and found these, both in loops that increment pointers by 32.

MCP :

• ldr x2, [x0]
• str x2, [x1]
  These instructions were executed in a loop

FCP :

• ldp q0, q1, [x0]
• stp q0, q1, [x1]
  These instructions were executed in a loop

Can these two behave differently ? memory accessing and register used are different, what can I infer more from this ?

Both routines are for measuring the Bandwidth, wanted to understand the difference. Tried going through source code.