3dfx Interactive was a company which specialized in the manufacturing of 3D graphics cards and graphics processing units. Initially dominating the field, by late 2000 it underwent one of the most high-profile demises in the history of the PC industry. It was headquartered in San Jose, California until it left the graphics business and its intellectual assets were acquired by its one-time rival, NVIDIA Corporation.

Early history

3dfx was formed in 1994, and two years later in 1996 it released its Voodoo Graphics chipset, later retroactively known as the Voodoo 1, after a massive drop in EDO RAM prices made such a card feasible for the mass market. The Voodoo was the first 3D accelerator that actually accelerated graphics performance, most systems of the era performed no better, if not worse, than software rendering engines.

In order to ensure the best performance, 3dfx developed their own Glide API for programming the card. Glide essentially exposed the internal hardware to software directly, as opposed to the hardware-neutral APIs of the era, OpenGL and QuickDraw 3D. At the time, 3DFX's main competition was from PowerVR, who produced a similar add-on card.

Much of the Voodoo's popularity was due to the close ties between Glide and OpenGL. This was because Quake used a cut-down version of OpenGL known as MiniGL that ran quickly on the Voodoo. The massive popularily of Quake led many gamers to buy a Voodoo for this game alone.

The Voodoo was notable because of its lack of 2D-display support; it functioned as an add-on card to an existing 2D video card, with the video output of the 2D card being piped into the Voodoo and then back out for display. Although this is a fairly annoying limitation, hard-core gamers were more than willing to put up with it for what was the ultimate in performance.

In August 1997, 3dfx released Voodoo Rush, combining a Voodoo chipset with a 2D chip from Alliance Semiconductor on the same circuit board. Unfortunately it performed worse than the Voodoo primarily owing to the fact that the 2D and 3D cores shared the same IRQs and memory. It was alsowas complex to make, and had poor quality 2D graphics. Later versions used a Cirrus Logic chip which solved most of the issues, but the Voodoo Rush never really took off and was dropped at the end of that year.

In 1998 they released Voodoo's successor, the Voodoo 2. This was generally the same as Voodoo, but it had a second texturing unit installed allowing two textures to be drawn in a single pass, a higher clock-rate, a wider memory bus (192-bit, compared to Voodoo's 128-bit), and a support for larger amounts of memory (up to 8 MB texture / 4 MB frame buffer compared to the Voodoo's 4 MB texture / 2 MB frame buffer) allowing a maximum resolution of 800×600 with higher quality textures. It also had the capability to work in Scan-Line Interleave (SLI) mode, in which two boards were connected together, each drawing half the lines of the display. SLI essentially doubled performance, and combining the two cards' frame buffer memory allowed resolution up to a then-impressive 1024×768. A problem with the Voodoo 2 was the fact that it required three chips whereas competing products such as the ATI Rage Pro and the NVIDIA Riva 128 were all single-chip products with integrated 2D cores.

Near the end of the 1998 3dfx released the Voodoo Banshee, which was basically a Voodoo 2 (with no multi-texturing on it) and 2D core (some parts made by Elpin Systems) integrated into a single chip. It was also clocked slightly higher than Voodoo 2, though it only had a 128-bit memory bus, like the first Voodoo. Because of its lack of the Voodoo 2's second texture unit, its performance was variable; in some situations it performed better than Voodoo 2, in others it performed more like Voodoo 1. While it was not a hit on the scale of Voodoo 1 or 2, it sold a respectable number of units (mainly to OEMs), and didn't flop on the scale that Voodoo Rush did.

Decline

In mid-1999 the Voodoo 3 was released, which was at heart a dual-core Voodoo 2 with Banshee's 2D core. It was a compelling solution, since an SLI-configured Voodoo 2 took up three slots, including the 2D card. However, given its design legacy it lacked support for several technologies that its competitors, ATI Technologies, Matrox and NVIDIA had since integrated, most notably 32-bit color support, followed by textures greater than 256*256 in size. Just prior to the launch of Voodoo 3, 3dfx bought out STB Systems , which was one of the main graphics-cards manufacturers at the time. It's generally thought that this move was one of the main contributors to 3dfx's downfall, since 3dfx did not sell any Voodoo 3, 4, or 5 chips to third party manufacturers, while NVIDIA were selling all of their processors through third-party card makers. The Voodoo 3 sold relatively well, but disappointingly compared to the first two models.

Their next (and as it would turn out, final) product was code-named Napalm. Originally, this was just a Voodoo 3 modified to support newer technologies and higher clock speeds, with performance estimated to be around the level of the NVIDIA TNT2. However, Napalm was delayed, and in the meantime NVIDIA brought out their GeForce chip, which shifted most of the computational work from the CPU to the graphics chip. Napalm would have been unable to compete with GeForce, so it was redesigned to support multiple chip configurations, like the Voodoo 2 had. The end-product was named VSA-100, which stood for Voodoo Scalable Architecture.

The two initial products were the Voodoo 4 4500 (single chip) and the Voodoo 5 5500 (dual chip), with a further two parts, the Voodoo 5 5000 (dual chip, but with a smaller frame buffer) and the Voodoo 5 6000 (quad chip) due to be launched later. But by the time the VSA-100 based cards made it to the market, the second-generation GeForce cards had arrived, which offered substantially better performance. By this point ATI had also released their Radeon line, which performed competitively with the GeForce 2 line. The only real advantage the Voodoo 5 5500 had over the GeForce 2 GTS or Radeon was that it had a better anti-aliasing implementation, and didn't lose as much performance when AA was enabled. Voodoo 4 4500 was beaten in almost all areas by the Geforce 2 MX and Radeon VE.

The Voodoo 5 6000 never got to the market, due to a severe bug resulting in data corruption on the AGP bus on certain boards, and was limited to AGP 2*, which would have prevented its use on the then-new Pentium 4 motherboards. Later tests proved that while the Voodoo 5 6000 would have been able to outperform the Geforce 2 GTS, it would've been outperformed by the Geforce 2 Ultra and the Geforce 3. The Voodoo 5 5000 never got launched either, as the smaller framebuffer didn't significantly reduce cost over the Voodoo 5 5500.

Voodoo 4 was as much of a disaster as Voodoo Rush, and while Voodoo 5's sales were respectable, they were nowhere near as good as 3dfx needed. In late 2000, several of 3dfx's creditors decided to initiate bankruptcy proceedings. 3dfx would have had virtually no chance of winning these proceedings, and instead opted to be bought by NVIDIA, ceasing to exist as a company. Most of the design team that were working on Rampage (the successor to the VSA-100 line) were transferred to the team working on what has since become the GeForce FX series.

3dfx's decline is a matter of debate, but it is most often attributed to managerial prioritizing of research and development. Voodoo cards were typically highly expensive, and left the mid and low end of the market to ATI and NVIDIA. NVIDIA chose short development cycles whereas 3dfx pursued lengthy development cycles, and NVIDIA and ATI cards had better overall performance, with Matrox holding the edge in image quality. The Rampage card, which 3dfx put much effort into but never was able to bring to market, is said to have been technologically several years ahead of the competition.

While some have theorized shipping the Rampage might have saved 3dfx, the fact remains the company never mastered designing cheap, high performance dies, with integrated high quality 2D acceleration, of the type NVIDIA pioneered. The success of Rampage world have depended not simply upon raw performance, but also upon costs of manufacturing. It is open to question, as to whether Rampage would have been a practical product, and whether it would have enabled 3dfx to retain a dominant high volume position in the graphics industry.

Chipset table

• *The VooDoo5 6000 was originally intended to have a core and memory clock of 183MHz, but all of the prototypes running at 183MHz stopped working after a short while. The only still-working VooDoo5 6000s all run at 166MHz, and 3dfx had decided to drop the 183/183MHz idea anyway.
• **Shared by two processors; effectively 32MB VRAM.
• ***Shared by four processors; effectively 32MB VRAM.