I genuinely can't understand the idiotic comments I've read all around the net about how Futuremark is paid by Nvidia and the whole Time Spy is an invalid benchmark, just because they didn't build it to perfectly fit GCN and nothing else, from what I've gathered? AFAIK, Futuremark's philosophy has all along been to demonstrate the way most games will be coded for the following couple of years. Not just for Time Spy but for their previous benchmarks as well. I have no trouble believing that Time Spy's async method could represent exactly that. In fact, for the next couple of years an async method that fits Maxwell (no performance penalty) and Pascal (minor performance gain) as well as possible is pretty much the first step of PC game optimization. That is something that MUST be done in order to not upset the majority of your client base (that is, the PC gaming client base). Yes, of course games should be optimized for GCN as well, but like it or not, for the PC version release that is a secondary concern. Console ports usually will have some GCN optimizations by default, but even those won't be 1:1 compatible with PC. PC optimization must be done in order to have a good PC version and Nvidia has the majority of PC gaming market for the next couple of years, how ever amazing products AMD will put out.
Seeing how Time Spy is NOT a console port, it doesn't need to be "paid by Nvidia" for the design choices to make sense. In general, PC games won't be tailor-made for GCN for a good while forward. Nvidia has a strong grip of the PC gaming market, and that grip in itself will affect the way PC games are made. No money or conspiracies needed for that to be true. Then, on top of all that, there are the actively protectionistic methods that Nvidia isn't shy to use. The latter, one could argue, are questionable but the former is the natural way of things. GCN is not the Only Right Way for gaming and indeed, it's not the most common one for PC gaming. It's only just beginning to really gain momentum. As such, I think Time Spy is pretty much what we should expect for the near future.