Well, you can try and run MySQL or Postgre off a RAM drive, but I suspect you won't get close to this result. And then your data is not on a reliable storage media.

Can't talk about MySQL or PostgreSQL, but with Oracle I am pretty sure you can get very much close to this result with cached data. And this is without using the in-memory option. However, talking about it withouth any experiments to prove a point is just speculation. Memory vs Disks changes a lot performance-wise.

It was expressly forbidden in the software contract to benchmark Oracle back in the 80s.

https://danluu.com/anon-benchmark/

It's forbidden to publish the benchmarks, not to make them. And it's because most users are rubes who can't tune 2 databases at the same time.

With the ClickHouse OLAP db I'm limited by memory bandwidth.

I believe you're conflating the name with the actual contents of the post. The demo uses the MemSQL columnstore, which is disk-based and leverages memory only for indexes, metadata, query processing etc.