The NV2 on the other hand, was never released, since it was a contract with Sega America to make a slightly more advanced version of its failed Sega Saturn, but the contract was not concluded due to the cancellation by Sega of the project , which was called the Sega Saturn V08.
The first two designs that NVIDIA made, were based on rendering by quads instead of triangles, so they did not follow the standard model when rendering scenes in 3D, this forced NVIDIA to follow the path of other manufacturers of graphics cards and it was with the NV3 that the NVIDIA legend began to take off with its first successful graphics chip.
The best NVIDIA graphics cards ever
Next we leave you a list, there are not all those that have appeared on the market, but, if those that marked a before and after in the graphics card market starting with the Pre-GeForce era until today, we hope you enjoy of this retrospective that we have prepared for you.
RIVA 128 (NV3)
Long before the GeForce, the Riva name was related to NVIDIA and its first major launch was the Riva 128, unlike its direct rival, the 3Dfx Voodoo Graphics, it is a graphics that integrated the 3D accelerator on a single chip and the VGA.
We are talking about a graphics chip that was created in the era where there were no shader units and they did not calculate the geometry on the graphics chip itself, so that stage depended on the CPU.
The Riva 128 despite being a notable card suffered from the poor quality of the DirectX API at that time compared to the proprietary Glide API of 3Dfx, but it is remembered for being the graphics card with which NVIDIA began its legend in the footsteps. false initials.
RIVA TNT Series (NV4 & NV5)
The second generation of the NVIDIA RIVA came with an improved raster unit that allowed them to take advantage of the greater power of the Pentium II and the AMD K6 of the time when it came to processing scene geometry.
Like the RIVA 128, we are facing a graphics chip that incorporated both the VGA part and the 3D accelerator, whose main novelty was its two texture units from which it owed its name, TwiN Texel .
It was the first time that NVIDIA surpassed 3Dfx in specifications, since it allowed rendering with 32 bits of color and supported textures of 2048 × 2028 pixels while the Voodoo were limited to 256 × 256. Also, the direct rival to the first Riva TNT, the 3Dfx Voodoo Banshee, had a single texture unit.
The Suicide of 3Dfx
NVIDIA became the leader in the graphics card market after the launch of the first GeForce, but to understand how NVIDIA came to lead the market and the fall of 3Dfx, the context of the time must be taken into account.
3Dfx made the mistake of purchasing graphics card manufacturer STB, to become a comprehensive graphics card producer . The 3Dfx move made card assemblers very angry. It is as if today NVIDIA or AMD decided to do without ASUS, Gigabyte, MSI, etc. To make their own graphics cards, which would cause a huge pushback from those manufacturers.
The 3Dfx movement made NVIDIA the apple of the eye of graphics card assemblers.This added to the fact that the DirectX 7 API was superior to Glide, led to the different studios that developed games on PC made the leap of the API. proprietary 3Dfx, Glide, to DirectX. This greatly benefited NVIDIA, which saw the PC gaming market stop buying 3Dfx cards to buy those with NVIDIA chipsets.
GeForce 256 DDR (NV10)
<![if !vml]><![endif]>
At the end of 1999, NVIDIA launched the first GPU in history, named for the fact that the entire graphical pipeline was executed on the GPU without the need for the CPU, this thanks to its T&L unit that was in charge of the calculations in the stages prior to rasterization, in which the geometry of the scene is calculated , and which were executed by the CPU up to that moment.
NVIDIA surprised everyone with the GeForce 256, and it was the graphics card with which it completely destroyed 3Dfx in the market, since while the Voodoo 4 and Voodoo 5 had prepared to compete against the Riva TNT 2, in NVIDIA they gave a giant leap with this card, which did not include two texture units but four, and they were the first graphics cards to include DDR memory.
Such was the dominance of the GeForce 256 and its successor, the GeForce 2, that several manufacturers left the graphics card market never to return and even its rival 3Dfx went bankrupt to ironically become part of NVIDIA. On the other hand, ATI became a rival of despite with the Radeon and the competition of these with the GeForce lasts today.
GeForce 3 (NV20)
<![if !vml]><![endif]>
The second turning point in NVIDIA history came with the launch of the GeForce 3.
In 2001 came the third generation of GeForce and with these the Vertex and Pixel Shaders , whose concept is the fact of being able to execute small programs on the GPU that allow manipulating vertices and pixels. Obviously the first implementation was very limited compared to what we have today, but, Shaders completely changed the way graphics are rendered and have been standard to this day.
The advantage of shaders is that without them if you want to implement a new graphic effect it is necessary to micro-wire it and therefore create a new chip, the shaders implemented for the first time in the GeForce 3 allow developers to apply effects on vertices and pixels through of programs that run internally on the GPUs.
GeForce 8800 GTX (G80)
<![if !vml]><![endif]>
At the end of 2006 the GeForce 8800 GTX appeared, based on the G80 chip, which was the first NVIDIA GPU to unify the units dedicated to shaders . The concept is that instead of using different types of units for each type of shader as had been done up to that moment, we switched to using a general unit that can be used for any type of shader.
Today all the GPUs on the market have their shaders unified, from those used in smartphones to the most advanced workstation. It should be clarified that it was not the first domestic GPU to unify the shaders, since the first was the ATI Xenos from Xbox 360, but as for the PC, the GeForce 8800 GTX was the first.
On the other hand, it was the first time that the term CUDA core appeared to refer to the ALUs in FP32 within the GPU itself, a term that NVIDIA continues to use today and was the first architecture that NVIDIA oriented in the computing market. high perfomance.
GeForce GTX Titan (GK110)
<![if !vml]><![endif]>
The first NVIDIA Titan first appeared in history in 2013.
I inaugurate a new segment in the market for graphics cards for home computers, a much higher range than what we were used to up to that moment, designed for those who want to have graphics cards with the maximum performance regardless of the price.
When the GTX Titan was launched it seemed crazy to many, today with models like the RTX 3090 on the market it is more than clear that the range that NVIDIA started with the GeForce GTX Titan is in top shape.
The Titan range is one of NVIDIA’s flagship ranges for enthusiasts and that is why we decided to recall the first of the GeForce Titan.
GeForce GTX 1080 Ti (GP102)
<![if !vml]><![endif]>
The Pascal architecture resulted in a huge victory for NVIDIA over AMD, but its GTX 1080 Ti is a card that many of us will remember for the enormous performance it gave us in games.
It does not mark a before and after in the history of NVIDIA in terms of architecture, but its performance was so great for the time that many came to think that NVIDIA had released a card a later generation by mistake. Even some of the RTX 2000 models were outmatched in a multitude of games against this brown beast.
GeForce RTX 2080 Ti (TU102)
The RTX 2000s represent the biggest architectural change to NVIDIA GPUs in history since the 8800 GTX with the addition of Tensor Cores and RT Cores . Based on the TU102 chip, it has been the basis for later NVIDIA designs such as the current RTX 3000 and AMD RX 6000 from AMD.
Although Tensor Cores were added in Volta, this architecture never had a domestic version, so Turing was the first GeForce architecture to implement them, incorporating specialized units for artificial intelligence and allowing the use of GeForce in said field of the computing.
As for the RT Cores, they are used to calculate the intersection of the rays in Ray Tracing, which has allowed us to start bringing ray tracing little by little to the domestic market in which games have begun to implement them and they have become the de facto standard for years to come.
These have been the most notable NVIDIA graphics cards in history, and we are sure they will not be the last and we will see this list grow in the future.
