setsunatenshi said:
I'm trying hard not to insult here, so I'll attempt to stay positive and on the point...
Higher resolution assets require higher memory bandwidth. GDDR6 and HBM2 provide much higher bandwidth for the same amounts of RAM when compared to GDDR5. Also, the die sizes are smaller, meaning lower costs and lower TDP (more efficiency) for higher performance. At the same rate, whenever 4K+ resolutions will be used, an even higher memory bandwidth will be necessary, thus requiring yet another jump in RAM efficiency. |
Kid, as I said before. You don't know anything about economy and technology.
Consoles are not a PC. PS4 doesn't have 8GB GDRR5, it's actually has 16x0,5GB of unified memory. That's why the bandwidth is so high (176GB/s). 16GB of GDDR5 will be enough to run all games in 4K/60 if needed.
You are also completely ignoring the fact how difficult and costly it is to implement HBM. GDDR is just slapping modules on a PCB. HBM is using complicated interconnects and thru-silicon via's to connect delicate parts together. The kicker is that the company that can handle that isn't the same as the one that makes the HBM itself. So a high-end GPU has to go through 2-3 different companies possession which is a lot of incurred cost.
