Add new Go Benchmarks

[antoninbas]

A few new benchmarks, which can be useful when validating improvements to the implementation.

[codecov]

Codecov Report

Merging #329 (58dd8ff) into main (614e68c) will not change coverage.
The diff coverage is n/a.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #329   +/-   ##
=======================================
  Coverage   73.67%   73.67%           
=======================================
  Files          19       19           
  Lines        2800     2800           
=======================================
  Hits         2063     2063           
  Misses        572      572           
  Partials      165      165           

Flag	Coverage Δ
integration-tests	53.82% <ø> (ø)
unit-tests	72.53% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

[dreamtalen]

LGTM.
While trying this benchmark locally, I found the IPv4 AddRecord and AggregateMsg benchmarks are approximately 20% slower than the IPv6 one.
Have you also seen similar results Antonin?

BenchmarkAddRecordIPv4Addresses-10                     	  335524	      3345 ns/op
BenchmarkAddRecordIPv6Addresses-10                     	  404529	      2855 ns/op

BenchmarkAggregateMsgByFlowKey/ipv4-10         	   58531	     20586 ns/op
BenchmarkAggregateMsgByFlowKey/ipv6-10         	   82375	     15100 ns/op

[antoninbas]

@dreamtalen For BenchmarkAddRecord, I have also observed that. It's because the IPv4 benchmarks is not the same as the IPv6 one (more checks). Although I will update it so that they do the same. It doesn't really make sense to have them be different.

For BenchmarkAggregateMsgByFlowKey, the IPv4 one is faster for me, as expected.

BenchmarkAggregateMsgByFlowKey/ipv4-12         	  181294	      6341 ns/op	     696 B/op	      12 allocs/op
BenchmarkAggregateMsgByFlowKey/ipv6-12         	  164910	      7327 ns/op	     792 B/op	      13 allocs/op

[antoninbas]

@dreamtalen I unified the test code, PTAL

For BenchmarkAggregateMsgByFlowKey, you could give it another try (I recommend adding -test.benchmem as well). Your computer may have been running something else at the time.

[dreamtalen]

LGTM

[dreamtalen]

@dreamtalen I unified the test code, PTAL

For BenchmarkAggregateMsgByFlowKey, you could give it another try (I recommend adding -test.benchmem as well). Your computer may have been running something else at the time.

Thank you. As expected, the IPv4 one now performs slightly better in both benchmark tests.