This article was originally published in 2003 in three separate articles. The article was written by Joel Hruska, who was a contributing writer to Sudhian at the time.  The information has now been combined into one long article - resource if you will - on anything and everything you ever wanted to know about the greatest video card that never was. 

A Fallen Titan’s Final Glory, Part I:  Setting the Stage

The decline and collapse of 3dfx in the waning months of 2000 was an untimely end for the company that had virtually invented and popularized the consumer-level 3D accelerator.  Once considered a virtually-unassailable market giant, 3dfx’s fa硤e of invulnerability hid major stress points that, through a mixture of external competition and internal mistakes, ultimately fractured and brought the company down.  The collapse was so dramatic it caught the entire market by surprise; while 3dfx did appear to be in trouble, not many people thought the situation was as bad as the events of December 2000 showed them to be. 

The Voodoo5 6000 (for those six of you who don’t know) was 3dfx’s last major product, and, though it was never released, it was the last to be manufactured in any kind of volume.  From its inception it was an attention-grabber, though its mammoth physical design and external power source would attract as much criticism as praise throughout its short and unofficial lifetime.  Based on the VSA (Voodoo Scaling Architecture) 100, the V5 6K was the ultimate over-the-top product; a 128 meg card when 32 meg was standard and 64 extravagant, a 4 chip architecture when other cards used only one (two at the most) and, inevitably, a (proposed) price tag that could choke a donkey.

It’s not possible to discuss the V5 6000 (and the entire VSA-100 product series) without also discussing the market conditions and historical events (can events from seven years ago BE historical?) that led to the ultimate demise of 3dfx.  Pay attention?there will be a quiz later.

The Dawning of a New Era:  1996-1997

Prior to 1996 and the introduction of the 3dfx Voodoo architecture, the world of computer gaming was a vastly different place.  Although primitive 3D environments were used in some games, even the fastest modern processors were unable to render these environments in high resolutions (for the time) while maintaining acceptable frame rates.  Even Quake, which was id’s first 3D game that allowed for jumping, climbing, and point-of-view rotation, couldn’t be run in a basic 640x480 resolution without looking like a slideshow.

This was a situation that benefited CPU giant Intel immensely.  With game performance dependent on processor performance, Intel could count on a steady crop of gamers buying or upgrading CPU’s, especially as games continued to demand more and more power in order to run smoothly.  Computers that’d whizzed by on Doom only a few years earlier choked on Quake; locking gamers into an upgrade cycle that distinctly benefited Santa Clara?until 3dfx introduced the original Voodoo architecture and torpedoed Intel’s carefully laid plans.

The original MMX instruction set (which boosted multimedia and video performance) was partially designed to increase game performance (and further boost sales).  By providing CPU’s with support for instruction sets that would increase 3D and multimedia performance Intel was both distancing itself from competitors like Cyrix, IDT, and AMD by offering features these companies did not possess, and introducing a feature set for which it could charge additional premiums.  MMX might’ve been only the first in a long line of instruction sets designed to boost 3D performance (and Intel profits)?if 3dfx hadn’t come along.

The Voodoo wasn’t the first 3D accelerator on the market, but it was the first 3D card on the market worth using.  The S3 Virge, in fact, performed so badly that it earned itself the nickname 3D "Decelerator" while Rendition’s Verite technology offered competitive video performance to a bowl of lukewarm tapioca?some of the time.  The Voodoo, in contrast, was fast, powerful, packed a whopping 4 meg of RAM on board, and was offered as an add-on card to a user’s primary video adapter.  Since many users invested in high-end 2D video cards at that time, releasing the Voodoo as a 3D add-in card actually made a great deal of sense, and didn’t force 3dfx to incur additional expenses by building a capable 2D engine or force users to choose between top-of-the-line 2D and gaming 3D performance.

For readers wondering just how much of a performance difference an original Voodoo could make we’ve included numbers gleaned from older sites where data is still available on GLQuake vs. ordinary Quake performance.  A Pentium 200 MMX system averaging a frame rate of 41 fps in Timedemo1 in Quake 1 (run in 320x200) rockets to 70 fps in GLQuake when a Voodoo card is used, for a whopping 70% boost in performance.  In 512x384 (probably the highest marginally playable resolution for this time period) we see Quake running at 21.2 fps on the Pentium 200 in software mode.  When we introduce the Voodoo we see performance jump no less then 300%, topping out an amazing 62 fps. 

Given these kind of results it’s easy to see why Voodoo cards sold hand over fist?and why Intel would have reason to be less than happy about these developments.  Although CPU power would continue to be an important factor in 3D performance, the emergence of the Voodoo knocked it firmly into second place when it came to determining just how fast a system would run in a given game. 

1996 and 1997 saw the mass adoption of the Voodoo, the introduction of 3dfx’s proprietary 3D API Glide (which came very close to dominating the market) and the establishment of 3dfx as *the* premiere 3D accelerator.  Most of 3dfx’s competitors (Matrox, ATI, S3, and Rendition) hadn’t really fielded seriously competitive products.  NVIDIA’s RIVA128 came on the scene in the fall of 1997 and became a popular low-cost option for OEMs, though it failed to make much of a dent in the higher-cost video market.

In February of 1998 NVIDIA launched the RIVA 128ZX, an upgraded version of the original with an 8 meg frame buffer.  The card took two slot spaces to install, was exceptionally loud, and offered only mediocre performance compared to its?.whoops.  Wrong card.  The RIVA 128ZX didn’t have many improvements beyond a larger frame buffer, but NVIDIA did have a full OpenGL ICD ready to go at launch.

3dfx did attempt to go after the 2D / 3D market (which would’ve put them in more-direct competition with NVIDIA), but their product of choice was the ill-fated Voodoo Rush.  Based on an out-sourced 2D design coupled with 3dfx Voodoo technology the card was plagued with compatibility problems and ultimately failed to have much market impact. 

The Storm Breaks

The Storm Breaks:  Spring 2000 - Fall 2000

The spring of 2000 brought the long-awaited launch of the VSA-100 product line and NVIDIA's GeForce2 technology, more or less simultaneously and the Radeon following a few months later.  The three are very different approaches to the same core principles, so we'll take a moment here to describe them all.

NVIDIA's GeForce2 GTS (Giga Texel Shader)

The GF2 GTS was built on the foundation of GeForce256 technology but added several additional features and abilities.  Like the GF SDR, the GTS used four rendering pipelines but added an additional texture unit to each.  This gave the GTS the ability to render eight pixels per clock and dramatically boosted its fill rate over earlier revisions of the card.  The phrase "Giga Texel Shader" referred to the card's 1600 Megatexels per second / 1.6 Gigatexels / second fill rate.

NVIDIA also added what they called a Shading Rasterizer whose function was to enable a number of pixel shading effects to be handled by the hardware itself.  Finally, the GTS offered an improved hardware T&L engine capable of processing up to 25 million triangles per second. 

The GeForce2 line was an enhancement over the GeForce, building on the latter's use of DDR RAM and offering several core enhancements, but at its heart the architecture was still a brute-force design that would later be shown to have its performance directly tied to the amount of memory bandwidth available.  Overclocking GeForce2-based cards was extremely popular as it was sometimes possible to achieve a 1-to-1 performance increase--overclocking the card 10% boosted performance by 10%, and so on.

Because we'll also be comparing the V5 line to the Radeon 64 / Radeon 7500, lets take a moment to discuss these products as well:

ATI's Radeon Architecture:

Prior to the Radeon, ATI wasn't considered a serious player in the 3D market.  Its products were late to market, its last serious attempt to enter the industry (the Rage Fury MAXX) had suffered from lackluster performance and the company's drivers were nothing short of abysmal.  The Radeon architecture launched a few months after the Voodoo5 did, and thus is considered here.

The Radeon had a hardware T&L solution (named the Charisma Engine) that was more advanced then that available in the GeForce2 GTS, as well as the industry's first bandwidth-saving memory architecture.  Unlike both the Voodoo5 and GeForce2 designs (which are both brute-force solutions), ATI developed a more elegant paradigm to use bandwidth more efficiently, which they named HyperZ.  Finally, the Radeon included its own shader capability, nicknamed Pixel Tapestry.  The card was a worthy competitor for the GeForce2, with its more-efficient design making up for a weaker fill rate. (The Radeon uses a 2x3 pipeline design, vs. the GeForce2's 4x2 and the Voodoo5 5500's 2(2x1) design. 

The Radeon 7500, also tested here, is nothing but a faster version of the original Radeon.  It includes none of the additional enhancements or features found in the Radeon 8500 design.

3dfx's Voodoo5 5500 / VSA-100 Architecture:

3dfx's VSA-100 (Voodoo Scaling Architecture) was a design meant to offer high performance via increasingly parallel operation rather than depending on a higher-and-higher clocked single GPU.  The first line of products, code-named Napalm, was the last iteration of the original Voodoo design, and was to be followed by Rampage, an all new product that incorporated, among other things, a hardware T&L engine.  3dfx designed a full line of products around the Napalm iteration of VSA-100, including:

Voodoo4 4500:  3dfx's new budget part, this card's main attraction was to be its 32-bit color and support for larger texture maps along with a small increase in speed over the Voodoo3.  16-32 meg versions of the card were standard, though its always possible that an 8 meg OEM-only version of the card might've eventually been produced.

Voodoo5 5000:  The V5 5000 used two VSA-100 chips for double the theoretical fill rate and memory bandwidth of the V4.  This would've been a 32 meg solution, offering performance above the V4 in most cases, though the 16 meg limitation on each chip could've caused problems at higher detail levels.

Voodoo5 5500:  The V5 5500 was 3dfx's last mid-range part to make it to market (the 5000 was cancelled).  Like the 5000, the 5500 used two VSA-100 chips, but in its case each VSA-100 processor had access to 32 megabytes of memory rather than the 5000's 16.  Performance was double that of the 4500 (most of the time) and slightly better then NVIDIA's GeForce DDR.  This was the card that made running in 2x FSAA possible nearly across the board, at least in lower resolutions. 

Voodoo5 6000:  The V5 6000 was a jaw-dropping monster when it was announced, with a massive 128 meg of onboard memory, 4 VSA-100 processors, and a fill rate over 1300 Mpixels / second.  The card used an auxiliary power connector (a target of much criticism at the time) and carried an enormous $600 price tag.

In addition to introducing the VSA-100 architecture, the Voodoo4 / Voodoo5 were the first 3dfx cards to support 32-bit color / 32-bit textures, 2048x2048 texture maps, an eight-bit stencil buffer, and support for up to a 32-bit Z-buffer.  The other major feature of the card was its new T-buffer technology.  The T-buffer was designed four five major effects:  Full Screen Anti-Aliasing (FSAA), real-time motion blur, depth of field blur, soft shadows, and soft reflections.  Unfortunately, the only one of these features that would ever be adopted was the one 3dfx implemented fully in hardware and could be turned on in any game?FSAA.  The others, to the best of our knowledge, never materialized outside of 3dfx technology demos.

Due to component shortages, 3dfx launched the V5 5500 and V4 4500 against the GeForce2 GTS, leaving the V5 5000 and the V5 6000 on the drawing board.  With the appearance of 3dfx's new flagship products, however, it became apparent that the V5 5500 wasn't going to catapult 3dfx back into the performance lead.  While the V5 5.5K wasn't slow, it offered performance only marginally better then NVIDIA's GeForce DDR, generally failing to keep up with the GeForce2 GTS.  The Voodoo5 5500 was around $50 cheaper then the GTS and had an edge when it came to image quality (not to mention a far-superior FSAA technology) but for consumers who'd believed the VSA-100 technology would utterly bury anything NVIDIA had to offer, the truth came as something of a shock.

In the year since the release of the Voodoo3 expectations over what VSA-100 technology could / couldn't do or would / wouldn't offer had built to unrealistic levels, aided, to some degree, by 3dfx themselves.  Previews from that time show T-buffer demonstrations with 3dfx on record as announcing that FSAA would function with a very small to non-existent performance hit?a statement that was obviously untrue when compared to the heavy toll running in 2x or 4xFSAA required from a video card.  3dfx was also forced to scale back their design plans slightly?early versions of the Voodoo5 6000 and Voodoo5 5500 used 183 MHz SDRAM, but it proved too difficult / expensive to deploy at that time.  Finally, of course, the cards were nearly nine months late, which also damaged their chances when compared to NVIDIA's grueling six month product cycle. 

As the summer wore on rumors began to trickle out that 3dfx's Voodoo5 6000?the long awaited uber-card and hoped-for champion of the 3D accelerator industry was going to be cancelled due to component problems and parts shortages.  3dfx's financial status was slipping badly as well, a trend that would continue through Q2 and Q3 of 2000.  3dfx, in fact, was never in all that great a shape financially, even during its heyday back in 1998 and early 1999?a fact which surprised me when I dug around a bit. 

The Voodoo5 5500 did gain some market traction, but by late summer 2000 3dfx was being hammered with problems from multiple sources.  The company's finances were in bad shape, NVIDIA slugged them with a major lawsuit, factory yields were very low (around the order of 25%), and the Voodoo5 6000 was nowhere in sight.  A potential deal with Microsoft to provide video solutions for Xbox might've saved the company, but Microsoft set their eyes elsewhere, eventually picking NVIDIA for the system's graphics provider. 

3dfx continued work on Rampage (the next generation of VSA-100 technology) but the card would barely tape out before the end.  Rumors persist that DDR versions of the V4 4500 and V5 5500 were in the works but never materialized.  It's too bad?had 3dfx copied NVIDIA and used slower DDR memory for the V5 5500 instead of faster SDR, costs might've been lower while raising overall performance.  Indeed, a DDR-based Voodoo5 probably would've met or exceeded GeForce2 performance.  As things stood, 3dfx never offered any sort of boost to Voodoo4 or Voodoo5 performance?the Hidden Surface Removal technology offered in the last 3dfx driver update was too buggy for use, and only worked in Quake 3 anyway.

< previous - (4 of 21) - next >

---

• Setting the Stage (1 of 21)
• The Golden Era (2 of 21)
• The Throne Trembles (3 of 21)
• The Storm Breaks (4 of 21)
• A Titan Falls (5 of 21)
• The Voodoo 5 6000 (6 of 21)
• Our Card (7 of 21)
• Test Bed and Drivers (8 of 21)
• Benchmarking Methodology (9 of 21)
• 3dMark 2000 (10 of 21)
• 3dMark 2001 (11 of 21)
• Evolva (12 of 21)
• Expendable (13 of 21)
• D3D Performance (14 of 21)
• Quake 3 (15 of 21)
• Jedi Knight & Serious Sam (16 of 21)
• Analyzing the Numbers (17 of 21)
• Rampage = MIA (18 of 21)
• Did Banshee kill 3dfx? (19 of 21)
• Conspiracy Theories, cont. (20 of 21)
• The lasting impact of 3dfx (21 of 21)