Yes and No.

Each CU has his own scheduler (called the CU scheduler), that's true. But GCN has also another scheduler for the draw and command queues, added with GCN3 because the CU scheduler was so inefficient. You may know this under Async compute, as that's how AMD branded that second scheduler. However, it can't handle too many tasks at once, thus getting very inefficient with higher CU counts.

Add to this the fact that the GPU driver must handle it's own scheduler (Number 3), which is handled by the CPU, and you can probably see the problems AMD has with their GCN schedulers.

Long story short: CU scheduler dispatches the commands (wavefronts) between the shaders, Async compute scheduler is supposed fixing any holes in those schedule pipelines, and the driver scheduler sets the order of the commands in each shader. Confusingly complicated, isn't it?

Not true. As I said, the front-end can issue wavefronts at a far higher rate than the CUs can process them. That is NOT an issue and that applies for both the graphics pipeline and the async compute pipes. Believe me, I KNOW, I'm not guessing.

Just like with the NX, I kinda despise and avoid hardware rumor threads. All they ever do is get your hopes up too high. And if they're true, great. I know it's fun for some people to dream but not me.

I really just posted to say that, while I'm typically an early adopter, I'm in no rush for next gen. I guess I'm getting old but even watching DF videos in 4K, I just don't see the improvements like I used to.

Ok..... try this. Watch a video say. Fallout 4. Then watch oneof say Cyberpunk. You see a difernce there?

Now imagine that comenext gen evrything will not just lookbetter but have more atmosphere and detail than what yo saw in cyberpunk.

Those DF videos you are watching in 4k are just that... 4k videos of the same current gen games. 

Man, that might be very expensive, unless they willing to sell at a considerable loss, or AMD is doing black magic.
Maybe it is the dev kit specs, and the consumer one will be cheaper, less powerful, less memory.
Looking foward

Not that relevant for consoles as developers are more likely to leverage the low-level API's instead to extract more performance.

And Polaris isn't toe to toe with Fiji. I provided evidence for this.

A random youtube video, evidence does not make.

250mm2 includes the entire chip by the way. That includes the CPU, memory controller, caches, fixed function logic and so on... If you think you are going to fit 73 CU's into that... Well. Not sure what to tell you.

Are you suggesting that Navi will need less transistors per CU than Vega 7? Do you have evidence for this?

Source please.

The complete performance isn't amazing, whichever way you cut the cake. It's mid-range.

We are. Because the argument is where Navi falls in AMD's product stack and how it compares to other parts.

Actually both.
Because Polaris had an x90 part. Aka. RX 590.

Wide GDDR6 buses can also take up a fair amount of die space as well, granted not likely as much as HBM, but certainly a step up over 256-bit GDDR5 controllers.

Ryzen 3 is about 80mm2 for an 8-core complex at 7nm.
Jaguar is 24.8mm2 for an 8-core complex at 28nm.

That doesn't include things like I/O.

At the end of the day, CPU cores themselves tend to be relatively small anyway, it's everything tacked on to hide bandwidth and latency (I.E. Caches) that drives die sizes of CPU's up.

AMD will be required to sacrifice some of the "density" in order to reduce leakage and noise so they can dial up clock rates.

Navi is Graphics Core Next based... Graphics Core Next isn't really designed for any particular node, it started life out at 28nm remember, got shrunk to 14nm and then 7nm... All the while iterative updates were introduced into the core design to extract more performance.

Navi isn't a new design built from the ground up. That's the ultimate crux of the issue, it's just more Graphics Core Next, nothing special.

At the end of the day, regardless of GPU, AMD's designs are inefficient, old and outdated... And sadly that is going to reflect on the capabilities of the next-gen consoles. 

You are better off comparing the Playstation 4 Slim against the Pro, they are built at the same node.
The Pro doubles the Slims CU count, but die size increased by around 32mm2.

We can already get an idea of allot of the design principles of Navi on the already established general design principles of Graphics Core Next that date back to the launch Xbox One and Playstation 4.

Graphics Core Next already has a multitude of bottlenecks inherent in it's design, it's why it's a compute monster, but falters in gaming.
Blowing out functional units will simply compound those inherent issues.

In saying that, Graphics Core Next is also highly modular, AMD can pick and choose what parts of the chip to update and leave the rest identical, which reduces R&D and time to market.

But the main reason for why Graphics Core Next won't scale beyond 64 CU's is simple.
Load balancing. - Which also occurs on the front end as well... And because the Geometry and Raster Engines are tightly integrated in a multi shader-engine outlay design like Graphics Core Next... The screenspace is split between the 4 engines and their respective clusters evenly.

Imagine you are splitting your display into 4x parts, the load balancing needs to be dynamic, so those parts shift so that the load is equal across all 4x shader engines.

What happens if we add more Shader Engines? It reduces the utilization across all the CU's. - There is only so much Screenspace that you can dynamically allocate with usable work to keep the entire chip busy. - One of Graphics Core Next's largest issues is effective utilization... Which was a much larger issue with the prior Terascale designs, hence why AMD introduced VLIW4 to increase utilization.

That all sounds great until the other guy releases cheaper and gets significant market share.  Taking a loss on a $400 console is fine because it will sell gangbusters if you can transfer all your PS4 games over and your account since there are many that would stay loyal if they aren’t paying $100 more for a graphics difference they might not care about

It's isn't toe to toe , agree is debatable. But it's very close on performance level.

a random Youtuber which already got acknowledged by many tech sites and tech Youtuber and got many insider is better than you.

250 mm2 with 72 CU  is possible with that price . 

Do i have to find one by one thousand of article of people like u who doubt this, and then arguing one by one of your previous article , while this thread you already back and forward arguing on the same thing?

It's a mid range desktop gpu performance  that run on  low mobile GPU/APU  . Yes it's amazing!!!

Well,you could be right.  But, history shows us that a console that launches at a price higher than $399 will struggle.  Every single one that has launched with a price higher than that has had trouble, at least until they lowered the price.  

As for phones, most of them are subsidized, they're much more mass-market, and they do a lot more for the average person than a gaming system does.  So, I don't see them as a valid comparison.  

Again. Your opinion on this is ultimately irrelevant and really not up for debate, I have already provided enough evidence.
Fiji beats Polaris and not always by a small margin either.

Fiji is:
Ashes of the Singularity: 39% faster.
Grand Theft Auto 5: 50.5% faster.
F1 2018: 23.8% faster.
Total War: Warhammer 2: 35.7% faster.
FarCry 5: 22.9% faster.
Shadow of War: 19.5% faster.
Final Fantasy XV: 22.38% faster.
Battlefield 1: 27.2%

Those aren't just a "small margin".

https://www.anandtech.com/bench/product/2373?vs=2397
https://www.anandtech.com/bench/product/2300?vs=2396

The only time Fiji will loose to Polaris is when a game demands more than a 4GB framebuffer... Like most IdTech powered games as they leverage Megatextures/Virtual Textures, but to be fair even Pascal (I.E. Geforce 1060 3GB) will tank in those instances as well.

To put the performance delta in perspective.. We are probably looking at a performance gap roughly equivalent to Vega 56 and Vega 7. That's not insignificant.

Prove it.

Prove it.

Not really. It could be better.