So more PR and nothing solid...gothca.
Star Swarm demo is slated for
release this month on Steam.
And if the folks over at DICE can ever get a handle on their bugs (which incidentally are often DRIVER RELATED), well the BF4 Mantle patch should show up soon. Myself, I'd rather see DICE take their sweet time, and get their first Mantle-related patch right, rather than rush it. I'll probably never play BF4 though (FPS give me motion sickness), so my opinion isn't as relevant.
And the Star Swarm benches, I tend to take at face value, based on the personalities involved. Of course, that engine is many many months away from a game release. Brad Wardell is a huge proponent of multicore usage, as are the Civ 5 guys.
Thief's North American release date is set for February 25th, 2014.
http://www.primagames.com/games/thief/feature/games-2014-thief
Other stuff is further down the pipeline, but where we have Developers already commenting about their personal experience working with the engine, I think that's a little more meaningful than just a press release.
One of the Oxide guys in the q&a mentioned 'seeing' a texture size issue (one texture was way bigger than the others), which became apparent afterwords using AMD's Mantle tools, which he was able to fix/resize fairly quickly. Others mentioned that they were able to more easily see how their resources were being utilized/which threads where where, utilizing the Mantle tools, and hence were able to reshuffle things to take better advantage of the GPU and CPU resources available. I also had the impression from many of their comments that the AMD Mantle tools were fairly easy to pick up, at least from the four that were commenting the most. The guy at the left in the Q&A video didn't say much at all...
The Oxide guys are not only making good progress with Mantle, they are now incorporating things (cinematic level fx, motion blur, etc) that would (according to them) simply have been too expensive resource wise with the current framework (DX11/GL).
The lifting of the Kaveri NDAs won't be all that meaningful, as pretty much all of the current benchmarking suites do not include Mantle (unless the BF4 Mantle patch is released by then). But Mantle isn't Kaveri specific in any case.
So yeah, until we see the releases, there is no way to really know for ourselves, but in the meantime, debating the merits of the moving parts is fun!
I fully expect Intel's Graphics division and NVidia to not stand still on this, and indeed they are not. The question is whether they can come up with a similarly performing/better solution, and how long it will take them to do so. Or perhaps jump on the Mantle bandwagon, which so far I've seen no direct indication of.
It's the 20+ titles onboard already thing that I think is the big news. In the past, when we've seen standards flounder (Glide, etc), it's been because multiple major developers didn't jump onboard. This time around, we already have 5 engines onboard, and it's been less than 6 months since Mantle's existence was officially disclosed. While it can take a year or more to bring a game to market, having this many onboard already is a good sign.
I certainly expect other developers to take the 'wait and see' approach, but I'm sure we will see more widespread adoption once Mantle hits the mainstream market.
The AMD guy in the Q&A did mention that AMD SPECIFICALLY limited the number of partners they brought onboard initially, so that they could stay focused on working with this small handful of developers firsthand. And based on the q&a, it sounded like the developers had good interactions with AMD on their end, although one commented that about 75% of the 'problems' which cropped up turned out to be on the developer's' end, and were solved within a day or two once they spotted the error/problem in their own code.
That revelation shows that even the early alpha version of Mantle was already very robust. In fairness to the developers, though, with Mantle they are doing some things in ways they had not tried/were not available before, so yeah there is going to be some hit and miss in the development process, as there always are.
So, the short form, soon... :biggrin: