OSDev.org

Whatever your scheduler is, someone will have a workload pattern that renders it wasteful.

Yes, this is certainly true. However, as was already pointed out, the real step forward of this paper is the techniques the authors developed for identifying and measuring such inefficiencies, which allows us (potentially) to identify common patterns that can be ameliorated, or even let us apply Amdahl's Law to decide which patterns are significant enough to address at what cost in development and added overhead (though applying that rule to software development is problematic, as the time and cost of software projects tend to be non-linear, even mathematically chaotic).