Update benchmark for R-arrow 20.0.0.2 on c6id.metal

[amoeba]

Hello! In Arrow 20, some notable improvements were made to our table join implementation in terms of speed and memory requirements so I wanted to update the benchmarks using the latest version.

I wasn't able to use the included scripts as-is to set up my EC2 instance but I think I was able to get close enough by referencing them. I also wasn't able to preview the results due to errors but I did poke around the CSVs and I think the we don't OOM on some of the questions we currently do. I'd love to see what the results look like now.

Let me know if I can add anything else here.

[Tmonster]

Hi Thanks for the PR!
I'm looking into fixing some other things on the benchmark, will take look at this today/thursday

[amoeba]

Thanks @Tmonster, much appreciated.

[Tmonster]

thanks! Will publish later today as well

[amoeba]

Awesome, thanks @Tmonster.