- Nov 26, 2021
- 270 (0.72/day)
- Mississauga, Canada
|Processor||Ryzen 7 5700X|
|Motherboard||ASUS TUF Gaming X570-PRO (WiFi 6)|
|Cooling||Noctua NH-C14S (two fans)|
|Memory||2x16GB DDR4 3200|
|Video Card(s)||Reference Vega 64|
|Storage||Intel 665p 1TB NVMe, Crucial MX300 1TB SATA, Samsung 830 256 GB SATA|
|Display(s)||Nixeus NX-EDG27, and Samsung S23A700|
|Case||Fractal Design R5|
|Power Supply||Seasonic PRIME TITANIUM 850W|
|VR HMD||Oculus Rift|
|Software||Windows 11 Pro, and Ubuntu 20.04|
Higher node costs are an important factor, but Nvidia's rapacity isn't solely due to that. AD103 is 295 mm^2, and in 2020, the similar N5 process was 82% more expensive than the N7 process used for Big Navi. The Navi 21 die is 520 mm^2. This means that TSMC would charge almost the same amount to manufacture AD103 and Navi 21. So the 4080 12 GB, even with good margins, could have been priced similarly to the 2020 MSRP of the 6800 XT.TSMC stated that the N5/N4 cost is about 2x higher than N7. And Samsung's 8N was cheaper than that (that was effectively a 10 nm class process).
As for your math, unfortunately it is not that simple. Process nodes have gotten so complicated, that the traditional nanometer metric does not apply anymore (with stuff like finfets, EUV, wider metal pitch and who knows what else, I do not even understand any of it). That is why they came up with those weird names.
TSMC's N5 has triple the density over Samsung's 8N. The N4 even slightly better than that. There is also a separate 4N process that NVIDIA is using here, and I do not know what the difference is compared to N4.
But AD102 has 2.7x more transistors with a slightly smaller die compared to GA102. The die size being so similar means that Ada chips are over 2x more expensive to make currently.
The cost increase is there, but their margins are another thing. As I said in another thread, maybe it is time to go back to making smaller and less power hungry dies.
Give up performance increases for one generation, instead utilize the efficiency of the new process and make the flagship model consume 250 W, sell it for $600-700 and call it a new generation of environment friendly technology. Would that not be a win in a world struggling for resources?
I think this is what AMD realized. I think they will let NVIDIA keep the pointless performance crown while they sell extremely efficient and affordable cards. I cannot wait to find out.