Queues: benchmark PIEO queues against all others

[polybeandip]

Do we have a working hardware implemenation of PIEOs? If so, HOORAY! Keep reading. If not, come back when we do.

This issue is the succesor to #2221 and #2276! In short, just as #2276 benchmarks the binary heaps against our specialized queues, we'd like to do a similar comparison for PIEOs.

More concretely, we want to do the following:

1. Implement two additional layers of logic on the PIEOs
   • strict
   • round_robin
2. Check that these agree 100% with the outputs of the specialized round_robin & strict queues!
   • For example, when fed rr_6flow_test.data, the PIEO implementation should produce rr_6flow_test.expect
3. Measure the performance! All the best hits from Synthesis 101: cycles, luts, worst_slack, registers, muxes, etc.

This last step should be as simple as simply running cycles.sh and resources.sh. However, plotting the data is trickier and will certainly require fiddling with plot.py (or writing another script).

Once you've done all that, don't forget to update cucapra/packet-scheduling#60