Add a benchmark which uses Google benchmark library

[egecetin]

~~Automatically generate performance benchmark information for website using Google Benchmark~~

Add a google benchmark based benchmark application

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 0% with 108 lines in your changes missing coverage. Please review.

Project coverage is 83.15%. Comparing base (573effc) to head (e8efae5). Report is 1 commits behind head on dev.

Files with missing lines	Patch %	Lines
...amples/PcapPlusPlus-benchmark/benchmark-google.cpp	0.00%	108 Missing :warning:

Additional details and impacted files

@@            Coverage Diff             @@
##              dev    #1392      +/-   ##
==========================================
- Coverage   83.35%   83.15%   -0.20%     
==========================================
  Files         276      277       +1     
  Lines       48086    48196     +110     
  Branches     9721     9697      -24     
==========================================
- Hits        40080    40077       -3     
- Misses       7115     7223     +108     
- Partials      891      896       +5     

Flag	Coverage Δ
fedora40	75.19% <0.00%> (-0.22%)	:arrow_down:
macos-13	80.65% <0.00%> (-0.21%)	:arrow_down:
macos-14	80.65% <0.00%> (-0.21%)	:arrow_down:
macos-15	80.62% <0.00%> (-0.22%)	:arrow_down:
mingw32	70.87% <ø> (-0.05%)	:arrow_down:
mingw64	70.88% <ø> (-0.01%)	:arrow_down:
npcap	85.22% <ø> (-0.09%)	:arrow_down:
rhel94	75.05% <0.00%> (-0.22%)	:arrow_down:
ubuntu2004	58.60% <0.00%> (-0.15%)	:arrow_down:
ubuntu2004-zstd	58.72% <0.00%> (-0.17%)	:arrow_down:
ubuntu2204	74.96% <0.00%> (-0.21%)	:arrow_down:
ubuntu2204-icpx	61.41% <ø> (+<0.01%)	:arrow_up:
ubuntu2404	75.22% <0.00%> (-0.23%)	:arrow_down:
unittest	83.15% <0.00%> (-0.20%)	:arrow_down:
windows-2019	85.33% <ø> (-0.01%)	:arrow_down:
windows-2022	85.35% <ø> (-0.02%)	:arrow_down:
winpcap	85.32% <ø> (ø)
xdp	50.50% <0.00%> (-0.15%)	:arrow_down:

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

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[tigercosmos]

Where can I see the benchmark result?

[egecetin]

@tigercosmos Check "Run benchmark" step https://github.com/seladb/PcapPlusPlus/actions/runs/10644856460/job/29510043289

Approx. 10Gbps parsing, 1Gbps crafting performance in GitHub CI (just looking bytes per sec). Also items per sec is a good indicator which can we tabulate as for mean packet sizes (For example, if avg packet length 64, 512, 1024, 1500 etc -> size* items per sec *8

[tigercosmos]

Will it show a warning if the benchmark performance decreases too much? Otherwise, people will not click into the CI and check the result.

[egecetin]

@tigercosmos Actually my aim is to generate and update the benchmark results automatically. I enabled PR runs for just testing. I'll disable it when the work is done, and the runs will be only from master or dev. Because benchmark results can be very flaky without a dedicated machine/runner unfortunately. This workflow and integration with Google benchmark are just for easier benchmarking in local PCs and to give an idea about to performance in PcapPlusPlus website (since current benchmark really old)

[seladb]

@egecetin I'm wondering if there's much value in publishing our benchmark without comparing PcapPlusPlus to other libraries. What do you think?

[seladb]

@seladb Actually it is just for give an idea about the performance in terms of packet per second and for easier local benchmarking (both disk and cpu sides). Github's hardware is good enough to get good results. This is why i added long sentence to README and a workflow to CI. I agree it alone does not have true meaning unfortunately so why I left the old benchmark also. I can (at least might try 🙂) also update to comparing with other libraries periodically. Maybe once a month or at least manually run after every release in a different repo.

Maybe what we really need is to add this as another example (and move the instructions to the internal README)? It's probably not worth running it in CI and uploading the results to the web-site repo

[egecetin]

@seladb It is ready to review if you have time