@ SiliconWars: you still don't get it. Any dip in the 30-60 fps range will result in smoother gameplay vs. not having the tech. And if framerate drops below 30fps, the gameplay will never be smooth one way or another.
You must be smoking something hallucinogenic to suggest that Mantle will magically cause fps to dip no lower than 60fps. As long as fps drops below 60fps, Mantle cannot resolve issues related to tearing/stuttering/lag that are fundamental to the monitor tech.
I am purely guessing here but doesn't this depend on the engine and the efficiency of the api, I mean if CPU bottleneck in a fx 8350 @stock + r9 290x @stock rig causing those dips, then mantle will help minimise those dips at the least right? 60fps is high reach accepted, but consider this if it does increases average fps, it may be doing so by increasing the dip point a few or many fps along with the same for max fps.
To my understanding the newest and baddest graphics cards on the both red and green do not sweat when it comes to an ever rising maximum fps roof threshold, to a point that it just gets ridiculous (for low res like 1080p - yeah FullHD is now the lowest res. ) so much so that 120Hz monitors feel really really smooth.
But on the other hand some games are punishing these cards left and right when it comes to the lowest fps margins. The developers of those games try to optimise the games for all settings by dealing with a black box api design-DirectX. Over the years they actually manage to formulate some tricks for general and optimisations specific to brands.
On the user side of things, both red and green teams generate profiles which is basically hardware specific tricks and optimisations developers woudn't know about and/or missed. To me this scenario is at least as exhausting as it sounded/guessed as to be when developing for different APIs, but then again that's why I said sounded or guessed as; because none of us
actually knows this even though some of us may be involved in the past or currently in development of games or programming. Because these people are not specificly involved in any phase of the current game development with mantle, and are just guessing while comparing it to glide times based on their experience
as users however!! so that they guess work is just that a
guess
Only people that can realistically give a comparison of
glide vs. mantle are developers that has worked/is working at both of these times.
Again only people that can realistically give a comparison of
Dx vs. Mantle are developers that has worked/is working at both of these times. So this rules out everybody except the actual developers (currently developing on mantle) that are giving statements as to how easy it is to develop for mantle.
One example is Nitrous,
they began developing for Dx in mind and multi threading application on it must be amasing to say the least looking at Dx portion of the comparison video. However, when they were producing and getting good results, they were introduced with mantle,
just one guy in rodeo hat take it up to him to change coding of their own engine into mantle ready one and it took
two months of that single guy to get the preliminary results. We all know that presentation now, and how they kept saying it is just bare minimum working version as it was just made to work in mantle nothing more (no optimisation..) and results are obvious.
Now you may argue that while Dx is actually great on it's own and mantle doesn't necessarily carry it's own merits, and used only for cinematic motion blur etc. You still don't know the full extents of it when that engine will be used in 3 games. IF dx is good performance vs mantle good performance at extra cinematic motion blur, then still it is a step in the right direction for a better IQ. IF dx is good performance at 2000 units vs mantle good performance at extra cinematic motion blur + 5000 and more units, then not only it is a step in right direction towards better IQ but also a
better experience.
And my guess is backed by yet another producer, Nixxes, which in these slides
http://www.slideshare.net/DevCentra...-new-amd-technology-by-jurjen-katsman#btnNext talks about both CPU and GPU performance improvement AND getting expensive compute based effects at low performance cost AND gettin a higher draw distance without losing performance.
Also on slide 26, it says and I quote "we spend a lot of time on consoles to develop best practices GCN and CPU: heavily multithreaded rendering, asynchronous compute, That's where the complexity is,
This now translates easily to the PC.