Quantcast
View Post
Pemalite said:

Can't think of any HDMI 2.1 devices that can do 4k 120Hz correctly. There are TVs and monitors that can do 4K 120hz, by using Y′CBCR with 4:2:2 or 4:2:0 subsampling... So I don't see the importance personally.

But no reason why they can't backport it to an older Graphics Core Next design anyway... Wouldn't be the first time AMD has done that. (Hence why some of their mobile chips had better video engines than the desktop GPU at some points!)

YCBCr colour space isn't optimal for interactive content such as games. You want full RGB and that especially applies in the case of HDR content as well .. 

HDMI is a hardware feature built into the GPU's display engine so it can't be backported unless a vendor overengineered their initial implementation but this is generally a bad idea since specs can change and it can also lead to interoperability issues as well ... 

The reason why some of AMD's mobile chips have better video engines than their desktop counterparts is down to the fact that they had an extra ~6 months to bake in more features ... 

Pemalite said:

AMD doesn't need to beat Intel, they just need to offer a good product with good performance at a good price with decent power characteristics and features.

Intels 6-core mobile chips absolutely dominate AMD in heavily threaded scenarios and will often turbo up higher for longer to boot.

Intels manufacturing is still relevant... Their successive process nodes are still on track for their targeted scheduled releases.

Doesn't need to beat Intel ? In an x86 market, it is a "winner-takes-all" competition. If Intel has just even a 5% performance advantage then they win by default! I don't think you realize the reality of just how cut throat competition can get ... 

If anything, today's reviews on Navi made a strong case for why AMD should wait on their Zen 2 APUs to integrate Navi since their RX 5700 XT is like for like to the Radeon VII in 1080p gaming. This sort of efficiency is very valuable in the portable space since lower resolutions are often encountered which makes Navi significantly more efficient even in comparison to the 7nm Vega. Integrating previous graphics architectures isn't ideal from a scalability standpoint because there's still an architectural sore spot regarding lower resolution graphics performance. Using less power and die area to achieve comparable performance at lower resolutions is optimal for the portable space ...

As for Intel's nodes being "on track", it definitely doesn't look like it from the outside perspective and we also have no idea if Intel are ever going to use smaller nodes for desktop CPUs ...