Editing Voodoo 5 (section)

==Architecture and performance==
[[Image:VSA-100 GPUs.jpg|thumb|250px|VSA-100 ICs, revision 220 and 320 shown]]
The VSA-100 graphics chip is a direct descendant of "Avenger", more commonly known as [[Voodoo3]]. It was built on a 250 [[nanometre|nm]] [[Semiconductor device fabrication|semiconductor manufacturing]] process, as with Voodoo3. However, the process was tweaked with a sixth metal layer to allow for better density and speed, and the transistors have a slightly shorter gate length and thinner [[gate oxide]]. VSA-100 has a transistor count of roughly 14 million, compared to Voodoo3's ~8 million. The chip has a larger [[texture cache]] than its predecessors and the data paths are 32 bits wide rather than 16-bit. Rendering calculations are 40 bits wide in VSA-100 but the [[operand]]s and results are stored as 32-bit.<ref name=wicher>[http://hothardware.com/Articles/3dfx-Interview-with-Peter-Wicher/ 3dfx Interview with Peter Wicher], Hot Hardware, December 15, 2001.</ref><ref name=avenger>{{cite web|url=https://www.techpowerup.com/gpu-specs/3dfx-avenger.g943|website=TechPowerUp|title=3dfx Avenger GPU Specs {{char|&#124;}} TechPowerUP GPU Database|accessdate=August 28, 2024}}</ref>

One of the design goals for the VSA-100 was scalability. The name of the chip is an abbreviation for "Voodoo Scalable Architecture." By using one or more VSA-100 chips on a board,<ref name=wicher/><ref name=vsa-100>{{cite web|url=https://www.techpowerup.com/gpu-specs/3dfx-vsa-100.g946|website=TechPowerUp|title=3dfx VSA-100 GPU Specs {{char|&#124;}} TechPowerUP GPU Database|accessdate=August 28, 2024}}</ref> the various market segments for graphics cards are satisfied with just a single graphics chip design. Theoretically, anywhere from 1 to 32{{citation needed|date=August 2024}} VSA-100 [[GPU]]s could be run in parallel on a single graphics card, and the [[fillrate]] of the card would increase proportionally. On cards with more than one VSA-100, the chips are linked using 3dfx's [[Scan-Line Interleave]] (SLI) technology.<ref name=AnandVoodoo5Review/><ref name=wicher/> A major drawback to this method of performance scaling is that various parts of hardware are needlessly duplicated on the cards and board complexity increases with each additional processor.

3dfx changed the rendering pipeline from one [[pixel pipeline]] with twin [[texture mapping unit]]s (Voodoo2/3) to a dual pixel pipeline design with one texture mapping unit on each.<ref name=FiringSquad>[http://www.firingsquad.com/hardware/v55500preview/default.asp 3dfx Voodoo 5 5500 Preview] {{Webarchive|url=https://web.archive.org/web/20090617235319/http://www.firingsquad.com/hardware/v55500preview/default.asp |date=2009-06-17 }}, Firing Squad, April 24, 2000.</ref> This design, commonly referred to as a 2×1 configuration, has an advantage over the prior 1×2 design with the ability to always output 2 [[pixel]]s and 2 [[Texel (graphics)|texels]] per clock instead of 1 pixel and 2 texels per clock.

This is the first 3dfx graphics chip to support full 32-bit [[color depth]] in 3D, compared to 16-bit color depth with all previous designs. The limitation of 256px × 256px maximum texture dimensions was also addressed and VSA-100 can use up to 2048px × 2048px textures. Additionally, 3dfx implemented the FXT1 and DXTC [[texture compression]] techniques.<ref name=FiringSquad />

The VSA-100 supports a hardware accumulation buffer, known as the "T-buffer". When rendering to the T-buffer, VSA-100 can store the combined outputs of several frames. This mechanism allows for creation of effects such as [[motion blur]] (if used temporally) and [[Spatial anti-aliasing|anti-aliasing]] (if used spatially).<ref name=FiringSquad /> VSA-100 supports rotated-grid super-sampling anti-aliasing (RGSS AA) modes, with a maximum anti-aliasing level determined by the number of VSA-100 chips in the SLI configuration. One chip allows 2× AA, two chips allows 4× AA, four chips provides for 8× AA and so on. The RGSS method of anti-aliasing combines multiple samples of each frame, resulting in higher quality than the brute force ordered-grid over-sampling of ImgTech [[PowerVR]], ATI [[Radeon R100|Radeon DDR]] and Nvidia [[GeForce 2 series|GeForce 2]].<ref>[http://www.firingsquad.com/hardware/multisamp/ Multi-Sampling Anti-Aliasing Explained], Firing Squad, February 13, 2001.</ref>

The chip implements a 128-bit [[synchronous dynamic random access memory|SDRAM]] interface,<ref name=4500agp>{{cite web|url=https://www.techpowerup.com/gpu-specs/voodoo4-4500-agp.c3548|website=TechPowerUp|title=3dfx Voodoo4 4500 AGP Specs {{char|&#124;}} TechPowerUP GPU Database|accessdate=August 28, 2024}}</ref><ref name=5500agp>{{cite web|url=https://www.techpowerup.com/gpu-specs/voodoo5-5500-agp.c3531|website=TechPowerUp|title=3dfx Voodoo5 5500 AGP Specs {{char|&#124;}} TechPowerUP GPU Database|accessdate=August 28, 2024}}</ref> again similar to the Voodoo3. Memory capacity and bandwidth is separately dedicated to each VSA-100 processor. While capacity is not cumulative across the entire card, bandwidth is effectively cumulative and thus a card with 2× VSA-100 processors has similar bandwidth to a single-chip graphics card using 128-bit DDR memory. Memory is clocked synchronously with the VSA-100 chip. Later, unreleased boards were planned to offer a 64-bit DDR memory design instead, in order to reduce board costs through lower complexity, while offering similar RAM performance.

Voodoo 4 and Voodoo 5 support [[MPEG-2]] [[video acceleration]].

While VSA-100 is an [[Accelerated Graphics Port|AGP]] 4× capable graphics processor,<ref name="falconfly1">{{cite web|url=http://www.falconfly.de/identify-3dfx.htm|title=Identify your 3dfx Hardware with Part No. and/or special Markings|website=FalconFly Central|archive-url=https://web.archive.org/web/20040607055946/http://www.falconfly.de/identify-3dfx.htm|archive-date=June 7, 2004}}</ref> 3dfx did not implement AGP texturing.

The Voodoo5 did not offer next-generation [[Microsoft Direct3D|DirectX 8.0]] [[Shader#Vertex_shaders|vertex]] and [[pixel shader]]s that would be found in the GeForce 3 and Radeon 8500, nor even [[Direct3D#Direct3D 7.0|DirectX 7]] features such as hardware transform and lighting acceleration for vertices (although T&L could be compensated by a fast CPU). Designed for scalability, it was nonetheless unrealistic for 3dfx to incrementally increase the number of Voodoo5 chips just to keep pace in performance with rival developments. Nvidia would further refine the GeForce 3's feature set into the GeForce 4 Ti (NV25) to satisfy multiple market segments (the Ti 4600 at $399 and Ti 4200 at $199), as did ATI which derived the lower-cost Radeon 9000 (RV200) from the Radeon 8500 (R100).