To put it mildly, the current video card market is going through not the best period. The evolution of the GPU industry was nothing like what we see today, and manufacturers have done a lot of experimentation to get to the current technology. PC Gamer Portal speakThe true origins of video cards and shares a selection of the strangest GPUs in history.
Is it true that NVIDIA invented the video card?
If you believe the manufacturer's own words, it all allegedly started in 1999, when the company invented the GPU and revolutionized the computer industry. But the reality is more prosaic. Graphics chips have been essential to the industry since the 1970s, long before NVIDIA began mass-producing graphics cards. Companies like RCA introduced their own chips at the time, and they sometimes produced incredible 64 x 128 resolutions.
In 1981, IBM introduced the Monochrome Display Adapter, used to display text and symbols. Hercules Computer Technology soon released its own video cards, and Intel joined them in 1983. That's right, Intel entered the GPU market much earlier than you might think.
But in the 1990s, there was a boom in video cards: it was a rather volatile period for the industry, because new companies and products appeared at the same speed as they disappeared. Therefore, in 1996, 3dfx Interactive released the Voodoo graphics chip, making real-time 3D graphics accessible to many consumer segments. However, such cards for 3D cannot handle 2D, which is why users have to buy a separate video card with a pass-through cable.
Meanwhile, ATI Technologies (later acquired by AMD) was also on the rise and launched its Rage line of graphics accelerators. Among them, notable is the Rage Pro from 1997 – an attempt to combine 2D and 3D solutions on one board.
Nvidia finally entered the market in 1993, but the manufacturer only became famous four years later with the launch of the Riva 128: a graphics card that accomplished what the Rage Pro could not. The manufacturer has combined three-dimensional acceleration with effective video and 2D acceleration, even at 100 MHz clock frequency.
In 1999, Nvidia introduced GeForce 256 and at the same time introduced the term GPU… But even here opinions differ somewhat. Sony was the first company to use this acronym five years before Nvidia when talking about PlayStation consoles, but in a very different context. Nvidia's definition of GPU includes processors for handling geometry, lighting, and rendering. Thus, Nvidia popularized and clearly defined the meaning of the term GPU, but did not invent the video card as such.
The strangest video cards in history
Nvidia FX 5800 Ultra
If you think your current gaming PC is noisy, it means you can't hear your old PC. The Ndivia FX5800 Ultra is called a “vacuum cleaner” and a “leaf blower” because the graphics card fan is so loud. By 2003 standards, the GPU was quite powerful, but unfortunately the manufacturer received countless complaints as soon as customers found out that they also needed earplugs to play comfortably on the PC. Even Nvidia's senior management laughed at this product.
3dfx Voodoo 5 6000
3dfx Interactive fell victim to the industry's tumultuous times in the 1990s and 2000s, but before going bankrupt and selling its remaining assets to Nvidia, it produced truly excellent graphics cards. The Voodoo 5 6000 never made it to the consumer market, but the GPU was phenomenal, to say the least.
The Voodoo 5 6000's gimmick is that it requires its own external power source. The manufacturer has squeezed up to four VSA-100s onto one board and connected them using the SLI method (no, not the method used by Nvidia). Each of these four devices offers 32 MB of memory at 166 MHz, as well as two pixel shaders (vertex shaders do not yet exist). A fully functional engineering prototype of the Voodoo 5 6000 sold at auction for $15,000.
Nvidia NV1
Nowadays everyone knows Nvidia as the GPU giant that controls the majority of the market, but even this company has had unsuccessful launches in the past. NV1 is one of them. An ambitious effort to bring general-purpose, multimedia GPUs to market. It handles 2D and 3D video, supports the Sega Saturn controller, and even has a built-in sound card. This is how Nvidia wanted to solve the problem of buying multiple expansion cards but ended up releasing a graphics card that was expensive, overly complicated, and often had compatibility issues.
The nail in the NV1's coffin was driven by Nvidia's desire to beat the clock. In an era where triangles were the standard for 3D rendering, the company released GPUs that used quadratic primitives. Unfortunately, soon after, Microsoft adopted triangular rendering as the standard for the DirectX API, leaving NV1 in an awkward position.
Gigabyte GV-3D1
Possibly the most impractical GPU in history. PC enthusiasts love to design their own machines, but with this graphics card it was necessary to build the computer exactly as Gigabyte required. On the other hand, the user only has half the video card.
It should be noted that the GV-3D1 is manufactured based on a proprietary, niche design. Gigabyte took two GeForce 6600 GTs and two memory modules, then put them on a board and connected them via SLI. At the time, dual graphics cards were not uncommon, but Gigabyte's approach limited the GV-3D1's compatibility to just one specific motherboard. Of course, the video card comes with it, but what if the buyer already has a motherboard? Sorry, nothing can be done.
To be fair, the GV-3D1 can be connected to other motherboards, but then only one of the two soldered 6600 GTs will fit in the slot, which kills the whole point of purchasing. Not to mention that it would be cheaper to buy the two 6600 GTs separately and then buy an SLI compatible motherboard to go with them.
