More testing on another computer, this time using Windows Performance Recorder (WPR), based on @toitoy’s suggestion.
It’s not very precise (lots of background processes that cannot be closed and other variables), but I saved the logs and can send taktik the files if he requests it. Not posting it since there’s quite a lot of sensitive information.
The values are weighted, so we can better see how they perform.
~
The setup:
Renoise config:
CPU overload prevention disabled so the playback does not stop on testing.
Now, for the (sorta) good stuff:
Windows Performance Analysis (WPA) for Renoise
The first plateau [1] in the graph above indicates Renoise opening. Then I wait a few seconds before hitting play [2].
After hitting play, CPU spikes. The crackling starts almost immediatly, very subtle, but quickly gets worse, until getting more subtle again. Upon note release and new notes attack phase there’s another intense spike [3] with even more crackling. Again, it normalizes and crackling is reduced considerably, yet still present. The spikes repeat on new releases/attacks.
This repeats 3x in the test. Then I hit stop once.
It gets stable quickly [4], similar to when I booted Renoise. I hit play again, more spikes and crackling [5], similar to [2]. My naive guess is that the release phase somehow spikes the CPU?
~
Windows Performance Analysis (WPA) for Reaper
Audio output settings are replicated. WASAPI shared mode, 1024 etc.
Monitoring chain is offline to avoid interferences (not bypassed, offline). Anticipative FX processing is DISABLED for the test only (since Renoise doesn’t have a similar feature).
Also, please keep in mind: note duration is different in the Renoise project and Reaper project. For this reason, I made some changes (see my version here).
In Renoise, the chords play for 6 bars at 4/4 (4 LPB, 32+64 lines). I changed the Reaper project to reflect this. The original had notes every 4 bars, which caused more frequent CPU spikes.
Now, for the testing.
There’s some “idle” CPU usage upon startup [1]. I compared it to Renoise and it’s actually a bit higher!
Like in Renoise, I waited a few seconds before hitting play [2]. CPU spike on play, just like in Renoise. However, no crackling or audible artifacts. It plays very nicely.
CPU stabilizes and spikes again upon pattern restart and note release [3]. Still no crackling, even with this bigger spike. I let it happen 3x like in Renoise. I stop, wait a few seconds [4], then hit play again [5]. Spike, no crackling. I close Reaper.
TLDR:
CPU usage ‘pattern’ seems to be virtually the same in both Reaper and Renoise.
Upon playback, Renoise has constant crackling that gets more intense on note attack and release phases.
Reaper uses a bit more CPU (roughly 4% more), but has no crackling, distortion or any other artifacts. It does spike like Renoise, but not audibly.
~
Well… this leaves me with more questions than answers. 
I haven’t logged it yet, but the performance in both is a a bit better with FLStudio ASIO and ASIO4ALL. The crackling still happens in Renoise, though.
Again, this is not a scientific test, but a rather limited and somewhat subjective analysis. Still, I hope it’s useful somehow.