If they had an embarrassingly parallel test, Intel would be exclusively promoting that misleading number just like they do with Cinebench. Let's not create even more totally unrepresentative benchmarks.