It makes perfect sense.

Depends, I can point to a ton of games where the CPU is a bottleneck on the Xbox One and Playstation 4.
The CPU bottleneck will shift depending on the game itself and sometimes even the scene that is being displayed on the screen.

The CPU assists in preparing those frames you know.

The CPU assists at rendering in many game engines... It was common especially in the 7th gen.
Shall I point out the rendering techniques the CPU was doing?

That is a bold assertion.
I was obviously "dumbing down" my rhetoric to make it more palatable for less technical persons that frequent this forum, if you would like me to stop, I would be more than okay to oblige and start being more technically on point?

Yawn.
The CPU is responsible for more than that... And you should probably list them, otherwise it is a little hypocritical if you are going to complain about my statement not being fully fleshed out and you go and do the same.
 

It is a GPU and a CPU issue. - Sometimes even a RAM issue.

Bullshit it's not everything.
FLOPS or Single Precision Floating Point Operations... Is a Theoretical number.

By that admission alone, Flops is irrelevant... Not only are they irrelevant.. But Flops tells us absolutely nothing about the hardwares actual capability, it doesn't tell us the amount of bandwidth a chip has, it's geometry capabilities, it's texturing capabilities, whether it employs any culling to reduce processing load, whether it has various compression schemes like S3TC or Delta Colour Compression, it tells us nothing of it's quarter floating point/double floating point/integer capabilities... It tells us absolutely nothing.
It's just a theoretical number that is calculated by taking the number of pipelines * instructions per clock * clock.

********************

I will try and keep this as simple as possible... But lets take the Geforce 1030.

DDR4: 884.7Gflops.
GDDR5: 942.3Gflops.

That is a 6.5% difference in Gflops... And you said flops is everything.
And yet we get to the crux of the issue. Gflops doesn't tell us everything else about a GPU, only a theoretical component.
In short... The DDR4 version is often less than half the speed of the GDDR5 version.

But don't take my word for it: https://www.techspot.com/review/1658-geforce-gt-1030-abomination/

****************

Or hows about a different scenario? (There are so many examples I can do this all day.)

Hows about we grab the Terascale based Radeon 5870 that operates at 2.72 Teraflops? It should absolutely obliterate the Radeon 7850 that operates at 1.76 Teraflops, right? That's almost a Teraflops difference huh? Both AMD based.
And yet... Again... Flops is irrelevant as the Radeon 7850 often has a slight edge.
But don't take my word for it: https://www.anandtech.com/bench/product/511?vs=549

Do you want some more examples of how unimportant flops are? I mean, I haven't even started to compare nVidia against AMD yet. Flops is everything right?

1. You can't point out many cases on the PS4/XBO with CPU limitations to frames. Those "limitations" are on the AI and on other things you don't really see. You know nothing. Consoles are closed hardware, where the floating point is dead set for developers. You clearly don't understand my dam example, which is as simple as I can make for a noob. Bandwidth.

2. Why do you say dumb things like "The CPU assists in preparing those frames you know." Well obviously, if you read my post, you don't have to say anything like that. It's obvious I know this as I listed many processes the CPU handles in gaming. But when drawing 3D pipelines, a CPU hardly does any of it. How do you miss my point? I am baffled.

3. The CPU again, doesn't render very many frames. It's EXTREMELY poor at doing it, that is why it's the brain that sends information to the GPU. The CPU is also responsible for telling the GPU what to do. While the GPU job is rendering what we see on screen at any given time. Remember we are simplifying here. The CPU sees and describes the different objects on screen, their location and other things. That information is converted by said GPU. A GPU's cycle is far more demanding than a CPU's. This is where you see a "bottleneck". If a CPU can't tell what the GPU should do fast enough, workload is limited in a gaming production role.

4. Frames are NOT a RAM issue today. So many developers are lazy and don't want to load in textures, and other things properly. Poor optimization. Can't give you this point at all. This is painfully clear in console gaming. Moving on.

6. Again FLOP are a good indication of similar architecture. WHAT DO YOU NOT GET? And if they aren't the same arch, or close, it just isn't a 1:1 ratio. Again, WHAT DO YOU NOT GET? But it gives a very good estimate on the strength of a GPU. Different architectures work in a different way, that's why it can't be a 1:1 direct correlation, yet the Flops themselves aren't meaningless. Compare Flops of new arch to that of the older. Notice how Flops get higher all the time? So yeah, you can easily draw a hypothesis based on Flops, even a pretty damn accurate one once you take in a few other form factors. Moving on.

8. I'm done. I gain nothing from this. You compared the 5870 and 7850. It's like nothing I said was even processed. I found it funny when you referenced that they are both from the same company. Oh boy... The same company doesn't use the same chips forever and put on turbo and stickers to make it go faster. LOL!

Based on the rumors that PS5 is using either a monolithic APU or dedicated chiplets, what if PS and AMD combined the APU and chiplet idea?

Instead of having one big APU or many smaller dedicated chiplets linked with infinity fabric, what if they had two smaller APU 'chiplets' linked?

Since we know it has Ryzen and Navi, each APU 'chiplet' could have 4 Ryzen CPU cores, and maybe 36 Navi GPU CU's let's say (the same amount as the PS4 Pro GPU). Take both APU 'chiplets' and link them together with infinity fabric, and use an I/O die if necessary. This way you would end up with 8 CPU cores and 72 GPU CU's, and could do so much more cheaply because your yields should be remarkably higher due to having much smaller chiplet APU's instead of one big monolithic APU.

Maybe this isn't technically possible or just wouldn't make sense, as a monolithic APU or dedicated chiplet design would be better overall possibly?

Yea, large monolithic chips hopefully can be replaced by clusters of chiplets in the future as that would mean much less wasted silicon waver area and should result in better prices. The problem is the bandwith of Infinity Fabric (IF), which afaik isn't quite high enough to seemlessly link several GPU chiplets into a big GPU. Even if AMD has a much more advanced IF rdy to go it could be prohibitively costly, as they don't seem to use it in their upcoming Navi GPUs (the one Lisa showed on stage at Computex was monlithic).

Based on the rumors that PS5 is using either a monolithic APU or dedicated chiplets, what if PS and AMD combined the APU and chiplet idea?

Instead of having one big APU or many smaller dedicated chiplets linked with infinity fabric, what if they had two smaller APU 'chiplets' linked?

Since we know it has Ryzen and Navi, each APU 'chiplet' could have 4 Ryzen CPU cores, and maybe 36 Navi GPU CU's let's say (the same amount as the PS4 Pro GPU). Take both APU 'chiplets' and link them together with infinity fabric, and use an I/O die if necessary. This way you would end up with 8 CPU cores and 72 GPU CU's, and could do so much more cheaply because your yields should be remarkably higher due to having much smaller chiplet APU's instead of one big monolithic APU.

Maybe this isn't technically possible or just wouldn't make sense, as a monolithic APU or dedicated chiplet design would be better overall possibly?

Yea, large monolithic chips hopefully can be replaced by clusters of chiplets in the future as that would mean much less wasted silicon waver area and should result in better prices. The problem is the bandwith of Infinity Fabric (IF), which afaik isn't quite high enough to seemlessly link several GPU chiplets into a big GPU. Even if AMD has a much more advanced IF rdy to go it could be prohibitively costly, as they don't seem to use it in their upcoming Navi GPUs (the one Lisa showed on stage at Computex was monlithic).

There are other rumors at the moment that Navi 10 is still partially GCN and partially RDNA, and that Navi 20 next year will apparently become fully RDNA. With AMD mentioning that PS came into the picture after Navi had been in development, maybe the reason for the hybrid GCN/RDNA was for them? Since PS can add future advancements to their planned past/present GPU architecture using AMD's semi custom branch, maybe it's possible that PS5 could use smaller multiple APU chiplets based on Navi 10 that incorporate a Navi 20 advanced IF chiplet design?

Basically the rumors stated that anything can happen? You don't have much choice than those two options.

That is one possibility, but won't happen.

You will likely have the I/O, CPU chiplets, GPU chiplet... That way AMD can optimize the process for the chip designs better.

The infinity fabric likely doesn't have the appropriate bandwidth for GPU chiplets of 36+ NAVI CU's to work together.
100GB/s isn't enough... Especially as said GPU's will be accessing bandwidth multiples more than that to local RAM.

It is a good idea in theory though. But CPU's tend to be less bandwidth intensive than GPU's.

There are engineering ways around it though.
AMD could set up a direct high-bandwidth link from the GPU chiplets that avoids the infinity fabric direct to DRAM or another piece of logic and have them work together there... And use the Infinity fabric for local chip communication that tend to be smaller transfers.
But that approach would get complicated pretty quickly.

Either way, for the GPU... The monolithic die is here to stay for the immediate future for GPU's, AMD simply hasn't shown they have solved the limitations yet.

AMD is trying to distance itself from GCN. RDNA is very much built on the foundations of GCN... Just like VLIW4 was built on the foundations of VLIW5... Remember AMD was championing how Vega was one of the largest deviations of GCN at one point as well.