Review of the GeForce RTX 2080 Founders Edition.

ZAH Computers Tech Editorial

ZAH Computers Hardware Research Desk

In our upcoming review, we will examine the GeForce RTX 2080 Founders (reference) edition in detail over the course of 40 pages. Although it is the less powerful variant compared to the Ti model, it still delivers impressive performance, rivaling the 1080 Ti in various scenarios. Additionally, the RTX 2080 is equipped with tensor and ray tracing cores. We invite you to join us as we explore NVIDIA's GeForce RTX 2080. A significant amount of new technology has already been discussed, particularly regarding the Turing architecture of the latest GPUs, which represents a pivotal shift in the graphics card industry. In addition to the standard shading engine, NVIDIA has incorporated RT (Raytracing) cores and Tensor (AI) cores into these new GPUs, both of which are operational. Is Turing the beginning of a new era in gaming graphics that will last for the next two decades? The answer largely hinges on the adoption rate among software developers who create games. Currently, several RTX games are under development, and numerous announced titles are set to utilize deep learning DLSS through the Tensor cores. However, the primary focus of the new RTX series is Raytracing. Thus, we welcome you to an extensive series of RTX reviews. We will commence with the reference cards and subsequently explore the AIB cards, as NVIDIA has chosen to launch everything simultaneously for reasons unknown. To begin, we will provide a brief overview of the tests conducted in this article, followed by an exploration of the architecture, before delving into real-world testing. It is advisable to have a drink on hand, as these reference articles tend to be quite detailed, encompassing a wealth of information. Beginning on November 20, 2018, two new products will be introduced, with the GeForce RTX 2070 set to launch later in October. The GeForce RTX 2080 will be part of this graphics card series, although it will not feature the complete shader count and GPU found in the 2080 Ti. Instead, it utilizes a distinct GPU known as the TU104. This graphics card will be equipped with 2,944 active shader processors, also known as Stream or CUDA cores, operating on the TU104 GPU at a base clock speed of 1,515 MHz, with Boost frequencies reaching up to 1,800 MHz. The TU104 chip comprises six GPCs, 48 SMs, and eight 32-bit memory controllers, resulting in a total of 256 bits. Each SM is composed of 64 shader processors, a 256 KB register file, a 96 KB L1 data cache/shared memory cache, and four texture units. The complete TU104 chip contains 13.6 billion transistors, featuring 3,072 shader processors, 368 Tensor Cores, and 48 RT Cores. Thus, it is evident that the RTX 2080 is a reduced version. An x8 NVLink connection is enabled, providing 25 GB/sec of bandwidth in each direction, totaling 50 GB/sec. It is important to note that clock speeds and thermal design power (TDP) may vary, as AIB partners offer different factory-tweaked models. The GDDR6 memory will be linked to a 256-bit bus, and depending on the clock frequency, it can achieve up to 448 GB/sec. The starting price for AIB partner products will be, while the Founders editions will be priced. This product will have a TDP of 225 watts. Will the 2080 break through the FPS barrier? Let us explore this and also demonstrate some techniques related to Raytracing and DLSS.