AMD announced its new RDNA architecture powering the Radeon RX 5700 Navi GPUs alongside the 7nm Ryzen 3000 lineup at their Computex keynote. Although the star of the show was the Zen 2 core, the Radeon part did get a decent amount of attention, and unsurprisingly, many curious journalists were able to get their grubby hands on the Navi die for a more accurate analysis. The Navi GPU die has an area of approximately 255mm². This is almost half as large compared to the TU-106-400 chip powering the RTX 2070.
The GPU on the RTX 2070 is a massive 445mm², and even the GTX Turing die, the TU116 comes in at 284mm², making it larger than Navi despite being notably slower. This is because of two reasons:
- Firstly, the Turing RTX GPUs in addition to the regular shaders or CUDA Cores, pack the Tensor as well as the RTCores, greatly increasing the surface area of the die.
- Secondly, the Turing GPUs, both the GTX 16 series as well as the RTX 20 leverage the 12nm process, while Navi is based on TSMC’s 7nm node, making it a whole lot smaller as well as more efficient compared to the GeForce products.
Read more:
But rtx or gtx series will always win whatever what kind of technology imagine if nvidia had 7 nm tech then it will 10 x times more powerful then any amd card
It’s because RTX chips house Dedicated RT cores and it’s not small amount. Also Performance that Navi gives is still limited to gaming. For other purposes , RTX chips are reliable especially the CUDA compared to OpenCL.
Amd will always try to catch up with nvidia… And one day it will outrun ..that’s for sure i think