I wouldn't trust New Relic charts. Maybe that is the source of all this disinformation. Get in the actual code and time everything yourself. If PHP is actually slow, you should be able to isolate the problem, like a loop within a loop that's just bad coding. You can make any language go slow with bad code. Check out my blog post on this subject:

http://pjbrunet.com/microtimers-on-hooks-to-measure-wordpres...