According to the German version of the Tom Hardware (Igor Laboratory), Nvidia closes the chip production not A used in Turing GPUs architecture. The site reports that the company plans to complete manufacturing by the end of May and, from there, to market the RTX 20 (2060, 2070 and 2080) with the version the most powerful of the chip, which takes the "Codename" A.
Nvidia releases 3 demos with real-time Tracing Ray for GeForce RTX and GTX
To better understand the effect of this change, it is important to know the difference between the two chips. This practice of having two different versions of the same card is quite common in many Nvidia models. After the first performance tests carried out by the company, the cores with the lowest performance are identified for deactivation. This, in theory, causes the product to affect its final value to become more accessible. However, this practice also enables other manufacturers to sell over-locked models with a factory, known as OC-ed, for higher values, as in case of t EVGA Ultra XC Series.
Essentially, these models are distinguished by a letter, usually used in the letter A, to identify the most powerful model. In the case of video cards with Turing architecture the most powerful models with factory cover is TU106-400A-A1 and TU104-400A-A1. It is noted that boards are cheaper not A (without over-binding) on the models TU106-400-A1 and TU104-400-A1. What Nvidia is planning is to disable non-A chip production and, in June, Turing RTX would all come with OC-ed.
The chips are expected to have the new codenames TU104-410 a TU106-410 and the change in value of the RTX cards should be felt, as the most powerful GPUs replace the more responsible cards – right now, those with the performance.
Nvidia GeForce 430.53 – Beta to solve a problem with high CPU usage
It seems that Nvidia is trying to rush past the loss and wants to establish a more effective competition with AMD, which is about to launch its GPUs with Navi Architecture.