Advertisement
Advertisement

Rumor has it that Nvidia's AMP 2020 architecture will hit the market. The Americans should take advantage of Samsung's new 7 nm process, specifically the EUV process (extreme ultraviolet lithography), in which plasma lasers write the structures into the silicon. However, the first 7nm GPUs come from AMD [VIDEO]. According to the report by TweakTown, AMD should continue to prepare its navigation generation for this July, and more can be expected in the context of this year's CES.

AMD had already announced a keynote on the subject of 7 nm.

AMD's new launch

Recently, even reports were circulating that AMD could be much better off than planned in terms of sampling phase and board design, yet first products should not be released until the end of the second quarter of 2019.

Advertisement

The current RTX 2000 series could, therefore, have been the highest of the feelings for 2019. Of course, a new edition of the existing generation would also be conceivable, for example with adapted clock rates and process optimisations.

Geforce RTX 2060

The Geforce RTX 2060 was just officially announced - although you can hardly speak of the "middle class" in the face of 369 euros. The new Nvidia model settles on the performance level of the GTX 1070 Ti, which in turn corresponds to a factory overclocked Radeon RX Vega 56, according to official figures.

After the launch of the RTX 2080 and 2080 Ti in September 2018, the RTX 2070 followed in mid-October. This graphics card was released in several versions, in addition to some heavily overclocked manufacturer versions with thick heat sinks, there are also much cheaper variants of this graphics card, often with a little less careful sorted chip (TU106-400).

Advertisement

Nvidia also offered a Founders Edition (FE) of the RTX 2070, which is compatible with the better selected TU106-400 and which, in contrast to the low-cost versions of the RTX 2070, also often come up with a slightly higher power limit and thus better performance. Nvidia did not send test samples of the RTX 2070 FE in Europe.

The latest member of the RTX series is based on the chip of the RTX 2070 [VIDEO], but has been slimmed down in some ways and now carries - at least in the case of the RTX 2060 Founders Edition - the identifier TU106-300. Instead of the physically existing 2,302 FP32 shader units of the RTX 2070, the RTX 2060 remain their 1,920. Notably, Nvidia has reduced the 8 GiByte of memory on the RTX 2060 to 6 GiByte by decommissioning two 32-bit controllers.

As for the physical hardware, the RTX 2060 Founders Edition is almost identical to an RTX 2070 Founders Edition - the same basic hardware with the same circuitry makes it possible. To our knowledge, Nvidia has taken no cost-saving measures on the in-house model, apart from the trimming of the GPU and the storage subsystem. Buyers of an RTX 2060 FE will receive the board and cooler design of the RTX 2070 FE.