In the final message of Nvidia's shareholders, CEO Jensen Huang was in the 12nm of his turring graphics, the match was "exceptional", compared to 7nm GP.
He points out that the tuning is much more effective than the speed 20 in 7 nm. Although the construction work of this ancient structure has been the most recent.
The CEO assures Nvidia, who works with TSMC, in designing its own process (Finite NVIDIA) and designing an efficient architecture designed for this particular process. Subsequently, he refers to using a generic 7nm process that does not bring good results and does not have a good effect. It is "Out of Scale Energy Efficiency," which helps NVidia win the competition. "
NVIDIA's CEO said it would be interesting to see how the Navy works in terms of energy efficiency, and AMD's current GPU is not very efficient. As GCN based navigation, it may not be effective in turing. In fact we have these suspicious sources.
What do you think about the energy efficiency of NVIDIA and AMD? Do you agree with Jansen Huang?
Source: alpha query