Leaked benchmarks for AMD Radeon RX 480 hits minimum VR spec for $199

[Alby_da_Wolf]

Pemalite said:

[...]

Alby_da_Wolf said:

Probably when the new production process will be better refined, and firmware and drivers too, increasing power efficiency, AMD will be able to increase RX 470 performances a little, and RX 480 a little more, to differentiate them more from each other and also to deliver a more significant increase from previous generation.

What production process? It's 14nm like the rest of the Polaris 10 lineup.

Alby_da_Wolf said:

About the 460, I expect from it the same, and more, like being used together with Zen CPU architecture to make APUs a lot more powerful than now, but still keeping power consumption within reasonable limit.

APU's aren't built around the concept of taking desktop GPU's and throwing them together with a CPU.
What AMD does is reserve a % of the die space in terms of transister counts and whacks a GPU that will "fit" into that space to meet various price/performance targets.

AMD *will* have Graphics Core Next 4 APU's, which is the same technology derived from Polaris/Radeon 480, but it will be scaled differently.
Not only that but they will be limited  by Dual Channel DDR configurations unless they allow for motherboard companies to implement a Sideport kind of concept with soldered memory on the motherboard dedicated to the GPU.
Otherwise a fast GPU is essentially wasted.

AMD has also given out plans to build a 300w~ APU with a ton of GPU hardware directed towards the HPC market, which should be faster than the Playstation 4.

Alby_da_Wolf said:

About drivers, AMD (but NVidia too with SLI) should work  to make CrossFire drivers more transparent, and allow game devs to use it as if the multiple GPUs were just a single, more powerful GPU, possibly requiring them to write specific code only if they want to go at a lower level and push the limits higher, but this should be considered necessary only for the most graphics-whorish titles. Surely mid-range CrossFire and SLI solutions, like using multiple cheap GPUs, or the most basic Hybrid CrossFire scheme, adding a single discrete GPU to an onboard or APU one, not being extreme configs shouldn't require specific coding to be used in games (and othe graphics SW).

Game developers can and do just that. AMD and nVidia cannot force developers to do anything, but they do work with all major developers and publishers to build Multi-GPU "profiles" for their drivers.
Pretty much every major PC release these days has support for Multiple GPU's.... Unless of course you buy your games from the Windows Store which are confined to the Universal Windows Platform. (Ugh.)

What you are describing about having two GPU's seen as one has actually been tried before, known as Lucid's "Hydra Engine" it was a dedicated chip on some motherboards... It even allowed you to pair an AMD GPU with an nVidia GPU, the way it works is it intercepts Direct X/OpenGL calls and divides the work up that way.
However it was buggy, had artifacts, less than optimal performance and it was expensive.

There are Pro's and Con's to AMD and nVidia's approach though, they can retain a degree of quality, fix issues and provide optimal performance, the downside is of course compatability, but nVidia and AMD are pretty proactive on that approach, I've been running Multi-GPU's since the late-90's, things are certainly fantastic compared to then, most games today do support Crossfire/SLI or will soon after a games release.

I think the next step to making Crossfire/SLI more feasible is to allow the community to build, test and optimize profiles for games and have the community vote for them and then propogated via the cloud to everyone.

 

About process, yes, I know, and I was meaning that maybe the whole range could still receive tweakings. We know that 480's power problem was of power management, taking too much power from PCIe instead of from auxiliary power connector, but also after fixing that, we see that power efficiency is quite lower than promised, and while it's quite normal that real world performance be lower than ideal ones, maybe we can still expect improvements not only from drivers and from improved board design, but also from the HW itself if the production process can still be tweaked.

About APUs, I wasn't precise, I should have written the use of the latest AMD GPU cores in them, and not literally putting a 460 in a Zen APU, but simply putting the GPU power of a 460, obtained with the latest cores available, in it. Yes, a sideport for GDDR could be a nice idea to avoid wasting the GPU power of the  higher-end APUs, although faster DDR4, and organised in quad- or even better eight-channel architecture, could at least mitigate the problem.

About drivers and games, thanks for all the explanations, my doubts come from the existence of games that cannot use more than one GPU, or cannot without giving more problems than benefits, maybe it could be possible to transparently offer them at least part of all the benefits offered by multiple GPU with APIs that keep where it should be stuff that is at a level low enough to be better managed by those that designed the HW than by game devs.