To be fair, scheduler misfeatures are a whole lot more interesting than benchmarks that scale in expected ways.

The title is misleading because the submitted link is to the third section of the review. A link to the full review was submitted last Friday [0].

[0] https://news.ycombinator.com/item?id=22266386

Computers exist to run software. If the computer doesn’t actually run the software well, then it’s not interesting that it seems powerful on paper.