Before the Nvidia GeForce RTX 2070 was released, it had a slightly different name, along with some significantly different specifications.
The Nvidia GeForce RTX 2070 maybe a couple of generations out of date now, but at the time of release, it was one of Nvidia’s best and most popular graphics cards and is still fairly widely used by PC gamers aiming to keep costs down. The card we got, however, was not the card that was originally envisioned, as seen by the emergence of a rare engineering sample prototype.
An engineer called Jiancheng Liu showed off the rare sample on Twitter/X, which revealed a number of differences compared to the card that was eventually released. The most visually notable was the use of the GTX prefix, indicating it predates the shift to the RTX branding used in the finalized GPU.
The real differences, however, are only apparent on the inside. The GTX 2070 engineering sample only has 2,176 CUDA cores, 128 less than the GeForce RTX 2070. The RTC 2070 uses the same TU106-400A-A1 chip as the RTX 2070 that saw release but with two disabled SMs.
Subscribe to our newsletter for the latest updates on Esports, Gaming and more.
By utilizing the 400A BIOS from an RTX 2070, Liu was able to successfully flash the vBIOS on the GTX 2070, allowing for a higher power ceiling and thus opening up overclocking options. This allowed him to achieve a performance improvement that equated to roughly 95% of what an RTX 2070 is capable of.
Nvidia has since discontinued all cards that used the GTX branding, but the reveal of this sample and its branding invites speculation. Were Nvidia planning on introducing reduced-spec GTX versions of its RTX 20 series cards? The removal of ‘super budget’ options from Nvidia’s GPU product lineup has left many PC gamers facing hefty bills when it comes to upgrading, perhaps keeping the GTX lines as a budget option might have been popular. Unfortunately, we will never know for sure.