Looking at openbenchmarking.org, it seems they are related to Phoronix?

As to why - essentially, Phoronix benchmarks are crap, for a number of reasons:

1. They don't provide things like standard deviation

2. They compile software in improper way

3. They don't even know what are they measuring in the first place

It's been pointed to them numerous times, and honestly, I think it's a waste of time.