add instruction trace sample

[derekbruening]

One common tool that people build is an instruction trace tool. While the existing memtrace tool also records instruction information for memory instructions, it does not record for other instructions.

We'd like to show how to gather an instruction trace without hitting prohibitive slowdowns from clean calls on every instruction (which can also hit tricky problems on ARM: xref #1698) or from calling a printf routine over and over (xref #1929).