Finer plot aggregation

[emilk]

What

• Workaround for Plot becomes jagged with min-max aggregation #4969

We would previously aggregate the plots over a time-window equal in width to a ui point.

The MinMax aggregator does a zig-zag between min and max, so this causes a very jagged look, especially on high-dpi screens.

With this PR, we switch to using a physical pixels as the default aggregation width, but for MinMax is is half a pixel. This reduces the jaggedness, but does not completely remove it. It comes at a cost though of higher tessellation cost.

I have an idea for how we can throw rayon on the tessellation problem though.

Before

After

Checklist

• I have read and agree to Contributor Guide and the Code of Conduct
• I've included a screenshot or gif (if applicable)
• I have tested the web demo (if applicable):
  • Using newly built examples: app.rerun.io
  • Using examples from latest main build: app.rerun.io
  • Using full set of examples from nightly build: app.rerun.io
• The PR title and labels are set such as to maximize their usefulness for the next release's CHANGELOG

• PR Build Summary
• Docs preview
• Examples preview
• Recent benchmark results
• Wasm size tracking

[github-actions]

Size changes

Name	main	4998/merge	Change
plots.rrd	194.56 kiB	1372.16 kiB	+605.26%

[nikolausWest]

Worth checking the before and after performance on the plot-dashboard benchmark script for this one

[teh-cmc]

lgtm as long as the impact on a plot dashboard is negligible

[emilk]

Test-data:

just rs-plot-dashboard --num-plots 10 --num-series-per-plot 5 --num-points-per-series 5000 --freq 1000 --save big-plot.rrd

	tessellation ms
main	2.9 ms
This PR, serial tessellation	4.5 ms
This PR, parallell tessellation	2.3 ms

Parallel tessellation is something I'm working on in a branch on egui.
Of course, it won't help on web.

[emilk]

Yeah, I think I will roll this back to using full pixel width, thinking especially about retina screens on web.