Earlier this week, subject matter publications were received for reviews of the NVIDIA video map GeForce RTX 4090 and unpacked them; however, they will only tell about the new features and features of their work in the coming week, closer to the official launch of the video cards on 12 October. However, the Internet resource Igors Lab has decided to clarify questions about the giant dimensions of new accelerators and the appearance of 600 W rumours of energy consumption.
The report states that during the design phase of the Ada Lovelace-based accelerators, the NVIDIA engineers were driven by the fact that the graphic processors would be produced on Samsung's 8 nm process. In this case, 600 W would be the exact amount of energy needed to increase the frequency and productivity. The design of the accelerator and all internal components were designed with an energy consumption of 600 W.
However, in the development process, it was decided to use new graphic processors produced by Taiwan's TSMC 5 nm capacity. This was a turning point because the developers were able to significantly reduce the accelerator's energy level by maintaining high productivity. Since the design of the product and internal components was coming to an end, the refrigeration elements were designed to operate with 600 W power, although in fact GeForce RTX 4090 consumes 450 W energy. This explains why the new video map has such impressive dimensions. It is likely that video maps consuming more than 450 W of energy will still be available on the market, as the cooling system reserves are sufficient to cope with this load.