I was expecting just the car/spaceship and a part of the track to have 8K textures whereas everything else that is within 20 meters will have 4K textures and from 20 - 50 meters to sport 2K textures. Everything past 50 meters will just have 1080p to 720p textures. 

really?

the amount of actual Megabytes of RAM that exist during the same time frame is superior with the eDRAM

on purpose, any luck why xbox strugles with 1080p?

even though i gave you a clue and even entioned your daddy sony?

Again you didn't even address what initial concerns I had within the first comment of this thread. 

That line was worthless and unnessecary as well. Another crappy post from a corporate apologist like yourself but at this point I'm not really surprised. 

User was banned for this post - Kantor

well, is not me you should get angry

is shinen, why dont you tweety them about what you think of their comments?

and why did you bring up haswell anyway?

didnt know that they also had the edram in the same die with the gpu or cpu

wii u has the edram inside the gpu, the only one that access it as cache is the cpu, (probably between 2 to 4 megabytes like the 476fp l3 cache)

besides, is not that ipressive, even gamecube had 512 bits for the ebedded memory

Once again I had already addressed the issue a few pages ago. Shin'en is obviously damage controlling much like how cevat yerli did it for the X1 when developing ryse son of rome. The one who should be angry is you and a few other nintendo apologists like yourself and eyeofcore because they were being partially dishonest about the capabilites of the WII U. 

dihonest h?

yea right

once again gamecube has schooled you

1024 bits is dishonest specially sinc gamecube had 512 bits and 360 4096 bits

ports are ipossible with edram of 1024 bits sinc gives you just 35gb/s

i can beat that with just 4 megabytes of the old embedded meory of gamecube

qalread not just shinen, but also renesas has pointed that the edram should be 8192 bits

sorry, but they are nmorentrustful than you

on purose, haswell coparasion is a joke, specially since haswell has the edra in a sparate die and not on the sae die like wii u does

haswell edram is limited by an external bus of 512 bits, wii u edra is not limited by such a thing

It doesn't matter whether the die is separate or not. You still don't understand the purpose of a cache! I wonder why you ain't answering my question as to why X1 mulitplat games are having trouble achieveing 1080p ? 

Thanks for giving the shittiest post on the planet. Why don't you take the fight to tomshardwareshare or anandtech ? Oh right, cause you'll get pwned badly like eyeofcore did LMAO. 

it doesn matter?

yea right

tell that to the old xbox, if it wasnrt in a separate die then there wouldn be no ecxternal bus 32gb/s limitation

i understant perfectly well what caxche is, what you dont undesrtand is that 32 egabytes is far ore than the tipiocall cache that standard pc gpus pack

i am not saying is not used as cache, just that the cache so big copared to the gpu standards like 8kb or 64kb for texture cache or global data share, etc that could be considered as the gpu ram

analogies dude, analogies

they have trouble achieving 1080p since using edram or esra is trickier than ram, bgesides, teh debelopers of ryse rome already said that are using the esram mnore for rich environents and other stuff using g buffers aned other stuff like msaa

if they were to remove those other buffers ande only use the esram forn the 1080p, then there would bge no such strugling,  of cours there would be sacrificies like less quality textures etc

we dont even know if they use compression for textures , if they dont they are using nto much emory for something they cold have used less

again, ask daddy sony why is trickier to use edram or esraj than main ram

For the record I don't contradict shinen or renesas, you just don't seem to understand that it's the internal bandwidth that's being discussed.
The external bandwidth, that connects the eDRAM to the GPU and/or CPU is substantually lower.

Allow me to break it down for you some more, again.

With the eDRAM you have a bunch of functional units, think of these as peoples houses, in order for all the functional units to talk to each other and pass on work you need a big fat and wide road to connect them all, that's where the internal bandwidth comes into play.
Typically, internal bandwidth is stupidly fast and low latency. (They come with different implementations such as a Ring or Hub bus etc')
However once it comes time for all the work the functional units have done to be passed onto the GPU it needs to exit a much much much slower toll booth on a smaller, skinnier road.

Once again another crappy post.