Optimize encode by avoiding list of field props

[ypconstante]

Today encode_with_message_props creates a list of FieldProps before calling encode_fields.
Removing this temporary list and :ok tuples on encode_field made encode faster and reduced memory usage.

I also fixed the bench script and updated benchee while workign on this.

Operating System: Linux
CPU Information: Intel(R) Core(TM) i7-10750H CPU @ 2.60GHz
Number of Available Cores: 12
Available memory: 31.17 GB
Elixir 1.18.3
Erlang 27.0.1
JIT enabled: true

##### With input google_message1_proto2 #####
Name                                              ips        average  deviation         median         99th %
encode (optimize-encode-b86fcba-encode)      152.30 K        6.57 μs  ±1020.36%        5.75 μs        8.90 μs
encode (main-d582a31-encode)                 127.27 K        7.86 μs   ±954.45%        6.78 μs       10.06 μs

Comparison:
encode (optimize-encode-b86fcba-encode)      152.30 K
encode (main-d582a31-encode)                 127.27 K - 1.20x slower +1.29 μs

Memory usage statistics:

Name                                       Memory usage
encode (optimize-encode-b86fcba-encode)         2.47 KB
encode (main-d582a31-encode)                    3.51 KB - 1.42x memory usage +1.04 KB

**All measurements for memory usage were the same**

##### With input google_message1_proto3 #####
Name                                              ips        average  deviation         median         99th %
encode (optimize-encode-b86fcba-encode)      142.43 K        7.02 μs   ±931.72%        6.29 μs        9.49 μs
encode (main-d582a31-encode)                 122.52 K        8.16 μs   ±877.64%        7.11 μs       10.84 μs

Comparison:
encode (optimize-encode-b86fcba-encode)      142.43 K
encode (main-d582a31-encode)                 122.52 K - 1.16x slower +1.14 μs

Memory usage statistics:

Name                                       Memory usage
encode (optimize-encode-b86fcba-encode)         2.28 KB
encode (main-d582a31-encode)                    3.31 KB - 1.45x memory usage +1.03 KB

**All measurements for memory usage were the same**

##### With input google_message2 #####
Name                                              ips        average  deviation         median         99th %
encode (optimize-encode-b86fcba-encode)      301.87 K        3.31 μs  ±2468.44%        2.76 μs        4.98 μs
encode (main-d582a31-encode)                 246.10 K        4.06 μs  ±2028.37%        3.35 μs        5.94 μs

Comparison:
encode (optimize-encode-b86fcba-encode)      301.87 K
encode (main-d582a31-encode)                 246.10 K - 1.23x slower +0.75 μs

Memory usage statistics:

Name                                       Memory usage
encode (optimize-encode-b86fcba-encode)         1.52 KB
encode (main-d582a31-encode)                    2.02 KB - 1.34x memory usage +0.51 KB

[whatyouhide]

Lovely, thank you @ypconstante 💟