Added a benchmark for game logic

[frutiemax]

Google benchmark parameters seem to work. Still need to figure out what to do with CPU iterations. The command to be called is openrct2 benchlogic xxx.sv6

[janisozaur]

Why can't it use the standard form

for (auto _ : state)
{
    context->GetGameState()->UpdateLogic();
}

?

[frutiemax]

Actually, you found the problem with my code, you actually need this. Why?

[chfast]

The loop is actually measuring the time, you don't need your timer. I'm just not sure it is safe to call context->GetGameState()->UpdateLogic() repeatedly in a loop.

[janisozaur]

Other than the comment I posted above:

1. Make sure you formatted your code
2. Make sure your commit message conforms to https://github.com/OpenRCT2/OpenRCT2/wiki/Commit-Messages

The xcode failure needs to be handled by someone with access to xcode. Ping @AaronVanGeffen @Gymnasiast

[janisozaur]

That's not what I meant. I think in current form it will cause problems because you enabled manual timing and provide your own measurements, while all this should be handled by state. Why is it necessary to do your own timing instead of using the benchmark's one?

[frutiemax]

It is not necessary to do manual timing. But it works nonetheless and it is one of the examples in the documentation of the Google Benchmark library.

[janisozaur]

Since it is not necessary, I would prefer to use the simpler invocation

[AaronVanGeffen]

@frutiemax I've rebased this PR and pushed a fix for the Xcode project. Do you intend to continue to work on this?

[chfast]

The loop is actually measuring the time, you don't need your timer. I'm just not sure it is safe to call context->GetGameState()->UpdateLogic() repeatedly in a loop.

[chfast]

You also don't need to call the function with BM_ prefix.

[AaronVanGeffen]

This is essentially doing the same thing as #10700. We need to decide which to merge.