Benchmarks results after all floating-point instructions implemented

[chfast]

I compared Fizzy 0.3.0 with #486 (which is the last FP instruction implemented).

The parsing time is unaffected (±2%).
The instantiation time is unaffected (±1%).

The comparison of execution times looks as flows:

fizzy/execute/blake2b/512_bytes_rounds_1                     +0.0377
fizzy/execute/blake2b/512_bytes_rounds_16                    +0.0389
fizzy/execute/ecpairing/onepoint                             (statistically not conclusive)
fizzy/execute/keccak256/512_bytes_rounds_1                   -0.0105
fizzy/execute/keccak256/512_bytes_rounds_16                  -0.0142
fizzy/execute/memset/256_bytes                               -0.0090
fizzy/execute/memset/60000_bytes                             -0.0068
fizzy/execute/mul256_opt0/input0                             +0.0398
fizzy/execute/mul256_opt0/input1                             +0.0440
fizzy/execute/sha1/512_bytes_rounds_1                        +0.0024
fizzy/execute/sha1/512_bytes_rounds_16                       +0.0049
fizzy/execute/sha256/512_bytes_rounds_1                      +0.0184
fizzy/execute/sha256/512_bytes_rounds_16                     +0.0249
fizzy/execute/micro/eli_interpreter/halt                     -0.0133
fizzy/execute/micro/eli_interpreter/exec105                  +0.0292
fizzy/execute/micro/factorial/10                             -0.0188
fizzy/execute/micro/factorial/20                             +0.0080
fizzy/execute/micro/fibonacci/24                             -0.0229
fizzy/execute/micro/host_adler32/1                           -0.0140
fizzy/execute/micro/host_adler32/100                         -0.0265
fizzy/execute/micro/host_adler32/1000                        -0.0283
fizzy/execute/micro/spinner/1                                -0.0359
fizzy/execute/micro/spinner/1000                             -0.0288

The mean is 1.4%, median is 1.8% (with micro benchmarks ignored).

I also compared the number of instruction in the execute() function (without the "cold" part):

0.3.0:   1517
all FPs: 2323  (+53%)