Add instruction-level profiling report

[morganthomas]

The profiling report provides instruction counts for each function. This is useful, but it would also be useful to have a profiling report which states how many times each instruction in the compiled code is executed. This would help when it comes to identifying hot spots inside functions. Such a report should be easily generated from the execution history. This instruction-level profiling report can be integrated with the object code pretty-printing, extended with #51, for easier reading.

[morganthomas]

I'm working on this.

[morganthomas]

The ExecutionHistory includes line-level profiling information, which is useful and should be added to the profiling output. Instruction-level profiling information would also be useful. To get that may require modifying execute_bytecode_helper. It seems that line-level profiling information can be output by extending execute_bytecode with a call to a new function which pretty-prints the ExecutionHistory.

[morganthomas]

Line by line cycle count pretty printing appears to be working as expected. Here is an example output for XMSS aggregation:

Line by line cycle counts
=========================

line 13: 1 cycles
line 14: 1 cycles
line 15: 1 cycles
line 16: 1 cycles
line 17: 1 cycles
line 18: 1 cycles
line 19: 1 cycles
line 20: 7498 cycles
line 21: 4000 cycles
line 23: 500 cycles
line 24: 0 cycles
line 26: 1000 cycles
line 27: 2000 cycles
line 30: 517 cycles
line 37: 2500 cycles
line 38: 500 cycles
line 39: 500 cycles
line 40: 500 cycles
line 44: 1000 cycles
line 45: 0 cycles
line 46: 500 cycles
line 47: 500 cycles
line 50: 3500 cycles
line 51: 7500 cycles
line 56: 33000 cycles
line 57: 36000 cycles
line 58: 36000 cycles
line 59: 36000 cycles
line 63: 3000 cycles
line 64: 3000 cycles
line 65: 3000 cycles
line 66: 3000 cycles
line 67: 3000 cycles
line 68: 3000 cycles
line 70: 3000 cycles
line 71: 3000 cycles
line 72: 3000 cycles
line 73: 60000 cycles
line 75: 6000 cycles
line 78: 1500 cycles
line 79: 500 cycles
line 80: 2500 cycles
line 81: 33000 cycles
line 86: 500 cycles
line 87: 500 cycles
line 88: 500 cycles
line 89: 33000 cycles
line 91: 500 cycles
line 93: 500 cycles
line 95: 0 cycles
line 96: 500 cycles
line 98: 500 cycles
line 99: 67000 cycles
line 101: 20355 cycles
line 102: 6785 cycles
line 103: 6785 cycles
line 104: 0 cycles
line 105: 0 cycles
line 106: 6785 cycles
line 107: 6785 cycles
line 110: 23508 cycles
line 111: 0 cycles
line 112: 7836 cycles
line 113: 7836 cycles
line 114: 7836 cycles
line 117: 26919 cycles
line 118: 8973 cycles
line 119: 8973 cycles
line 122: 31218 cycles
line 123: 10406 cycles
line 124: 10406 cycles
line 125: 10406 cycles
line 126: 10406 cycles
line 127: 10406 cycles
line 132: 1000 cycles
line 133: 0 cycles
line 135: 500 cycles
line 136: 64000 cycles
line 139: 2000 cycles
line 141: 500 cycles
line 142: 0 cycles
line 143: 1680 cycles
line 145: 1560 cycles
line 148: 1000 cycles
line 149: 75000 cycles
line 150: 53949 cycles
line 152: 46758 cycles
line 156: 2500 cycles
line 161: 0 cycles
line 162: 500 cycles
line 163: 500 cycles
line 164: 500 cycles
line 165: 1500 cycles

[morganthomas]

Closed by: #72