Heres some old press-releases and previews i dug up... see if you remember these, and if you do, join me in giggling at their success. 
Alliance Semiconductor Paladin:
<< SAN JOSE, Calif.--(BUSINESS WIRE)--May 5, 1998--Alliance Semiconductor Corporation (NASDAQ:ALSC - news) today announced its next-generation 3D graphics accelerator chip - the Paladin(TM).
This feature-rich 3D/2D/Video accelerator delivers photo-realistic 3D graphics by taking full advantage of the AGP high-speed interface with sideband addressing. The Paladin(TM) incorporates an advanced 3D set-up engine, a rich set of 3D features, 260MHz RAMDAC, and enhanced 2D and Video acceleration capabilities to fulfill the graphics quality and performance needs of consumer and business PCs.
The Paladin(TM) supports AGP interface in both 1X (66 MHz) and 2X (133 MHz) modes with full sideband addressing. Two variants of Paladin(TM) are available - Paladin(TM) LT with AGP 1X, and Paladin(TM) EX with AGP 2X. Both these versions also support PCI 2.1 interface at 33 and 66 MHz speeds. Paladin(TM) LT and Paladin(TM) EX are pin-compatible, and therefore allow manufacturers to address different market segments and price points with a common motherboard or add-in card.
The Paladin(TM) has a flexible memory interface, which supports EDO, SDRAM, and SGRAM up to 16 MB configuration. >>
Stellar "PixelSquirt":
<< The second 3D chipset, or 3D graphics core, which people probably haven't heard of before comes from a fairly new company called Stellar Semiconductor. They plan on releasing about 4 different chips based around a central design called PixelSquirt. The basic idea of pixel squirt is that unlike most current 3D accelerators out there right now, Stellar's products will be scanline renderers instead of polygon renderers or even the tile rendering systems of PowerVR. To understand how PixelSquirt works, it's probably easiest to consider it from the point-of-view of a tile based rendering system - except all the tiles happen to be 1x1 pixels.
Just like the PowerVR SG, the PixelSquirt engine makes use of deferred texturing and does all the hidden surface removal before the final pixel is actually rendered. Also, like the SG, the 'Squirt will allow for oversample based scene anti-aliasing, full support for all alpha blending modes, bilinear filtering and all the other usual effects one might expect. While PixelSquirt doesn't actually require a Z-buffer (to save on memory to keep costs down) it will still allow you to use a Z-buffer if you insist on it, to make use of any special effects that might specifically require Z-buffer lookups. PixelSquirt is designed from the beginning to be fully 32 bit, so image quality and image fidelity won't suffer even after a ton of alpha blends have been used. >>
Silicon Reality Taz3D:
<< Silicon Reality's first chip, the TAZ-3D has architectural roots stretching back as far as 1982. This, in a good sense, these guys know exactly what they're doing and have blessed the chip with some interesting more or less new features never seen before. Examples of these are single pass trilinear mip-mapping, giving excellent texture map quality, 9-bit digital to analog signal converter giving supreme gamma correction and 32 bit rendering pipeline. The chip also features three fog tables giving a possibility of many different fog types.
To this point it's all good, but nothing really special. The inevitable question of speed arises... In this section the TAZ-3D offers comparable specs to the PVRSG and MGA G200, 1.7 million polygons per second isn't anything revolutionary but should be able to kick at least some Voodoo2 ass. Pricewise the chip has got real potential. Costing merely 15$ per chip (in large quantities of purchase, but anyway..) cards powered by it can be expected to cost around 100$ >>
ARK Tiger3D:
<< ARK (Advanced Rendering Kernel) is a relatively unknown company founded in 1993 who have been delivering both relatively high-end and low-end 2D/3D chips for quite a while now but have now decided to give the hardcore gaming class 3D industry a try. The ARK Tiger 3D offers a performance similar to the PVRSG's and seems to have all the right features incorporated, although no fancy newbies like Anisotropic filtering or trilinear mip-mapping have been mentioned in the 3D spec. sheet. The 32bit rendering and 24bit Z-buffer promise great colors and the image quality of a decent 2:nd generation 3D chip. So at the price of around 100 bucks I don't see why this chip should/would fail. >>
Rendition RRedline (go here for more) :
<< Well, there is preliminary indication that the RRedline may achieve the current Holy Grail of the 3D industry, beating 2 Voodoo IIs in SLI configuration. In other words, this is going to be one fast chip. >>
Bitboys Glaze3D From an AnandTech interview thats here:
<< We have designed a completely new approach to 3D acceleration, Xtreme Bandwidth Architecture. XBA technology was invented during the design of our current 3D-graphics architecture. XBA consists of the eight-texel/four-pixel rendering pipeline, extremely wide 512-bit memory bus and our memory management logic. The memory management logic works as a highway system tying together the embedded DRAM memory, the external SDRAM memory, AGP memory and all units, which want access to the memory. A lot of the XBA inventions are behind this memory system, which handles 768 bits of data (eDRAM+AGP+SDRAM) every clock cycle, resulting in the massive bandwidth of 12.5 GB/s.
This huge bandwidth enables us to do everything in true color; we are not really interested in any 16-bit performance and dithered images. The bandwidth also allows us to do full-scene Anti-Aliasing with real supersampling. Although rendering game graphics at 1024x768 true-color resolutions requires around 2.5 GB/s of memory bandwidth, doing this with anti-alias enabled requires 10 GB/s of memory bandwidth! To say this in another way, products equipped with only external SDR or DDR memory are not able to do full-scene anti-aliasing at realistic speeds. The XBA architecture based products are really the first chips capable of improving the image quality with Anti-Aliasing, >>
Alliance Semiconductor Paladin:
<< SAN JOSE, Calif.--(BUSINESS WIRE)--May 5, 1998--Alliance Semiconductor Corporation (NASDAQ:ALSC - news) today announced its next-generation 3D graphics accelerator chip - the Paladin(TM).
This feature-rich 3D/2D/Video accelerator delivers photo-realistic 3D graphics by taking full advantage of the AGP high-speed interface with sideband addressing. The Paladin(TM) incorporates an advanced 3D set-up engine, a rich set of 3D features, 260MHz RAMDAC, and enhanced 2D and Video acceleration capabilities to fulfill the graphics quality and performance needs of consumer and business PCs.
The Paladin(TM) supports AGP interface in both 1X (66 MHz) and 2X (133 MHz) modes with full sideband addressing. Two variants of Paladin(TM) are available - Paladin(TM) LT with AGP 1X, and Paladin(TM) EX with AGP 2X. Both these versions also support PCI 2.1 interface at 33 and 66 MHz speeds. Paladin(TM) LT and Paladin(TM) EX are pin-compatible, and therefore allow manufacturers to address different market segments and price points with a common motherboard or add-in card.
The Paladin(TM) has a flexible memory interface, which supports EDO, SDRAM, and SGRAM up to 16 MB configuration. >>
Stellar "PixelSquirt":
<< The second 3D chipset, or 3D graphics core, which people probably haven't heard of before comes from a fairly new company called Stellar Semiconductor. They plan on releasing about 4 different chips based around a central design called PixelSquirt. The basic idea of pixel squirt is that unlike most current 3D accelerators out there right now, Stellar's products will be scanline renderers instead of polygon renderers or even the tile rendering systems of PowerVR. To understand how PixelSquirt works, it's probably easiest to consider it from the point-of-view of a tile based rendering system - except all the tiles happen to be 1x1 pixels.
Just like the PowerVR SG, the PixelSquirt engine makes use of deferred texturing and does all the hidden surface removal before the final pixel is actually rendered. Also, like the SG, the 'Squirt will allow for oversample based scene anti-aliasing, full support for all alpha blending modes, bilinear filtering and all the other usual effects one might expect. While PixelSquirt doesn't actually require a Z-buffer (to save on memory to keep costs down) it will still allow you to use a Z-buffer if you insist on it, to make use of any special effects that might specifically require Z-buffer lookups. PixelSquirt is designed from the beginning to be fully 32 bit, so image quality and image fidelity won't suffer even after a ton of alpha blends have been used. >>
Silicon Reality Taz3D:
<< Silicon Reality's first chip, the TAZ-3D has architectural roots stretching back as far as 1982. This, in a good sense, these guys know exactly what they're doing and have blessed the chip with some interesting more or less new features never seen before. Examples of these are single pass trilinear mip-mapping, giving excellent texture map quality, 9-bit digital to analog signal converter giving supreme gamma correction and 32 bit rendering pipeline. The chip also features three fog tables giving a possibility of many different fog types.
To this point it's all good, but nothing really special. The inevitable question of speed arises... In this section the TAZ-3D offers comparable specs to the PVRSG and MGA G200, 1.7 million polygons per second isn't anything revolutionary but should be able to kick at least some Voodoo2 ass. Pricewise the chip has got real potential. Costing merely 15$ per chip (in large quantities of purchase, but anyway..) cards powered by it can be expected to cost around 100$ >>
ARK Tiger3D:
<< ARK (Advanced Rendering Kernel) is a relatively unknown company founded in 1993 who have been delivering both relatively high-end and low-end 2D/3D chips for quite a while now but have now decided to give the hardcore gaming class 3D industry a try. The ARK Tiger 3D offers a performance similar to the PVRSG's and seems to have all the right features incorporated, although no fancy newbies like Anisotropic filtering or trilinear mip-mapping have been mentioned in the 3D spec. sheet. The 32bit rendering and 24bit Z-buffer promise great colors and the image quality of a decent 2:nd generation 3D chip. So at the price of around 100 bucks I don't see why this chip should/would fail. >>
Rendition RRedline (go here for more) :
<< Well, there is preliminary indication that the RRedline may achieve the current Holy Grail of the 3D industry, beating 2 Voodoo IIs in SLI configuration. In other words, this is going to be one fast chip. >>
Bitboys Glaze3D From an AnandTech interview thats here:
<< We have designed a completely new approach to 3D acceleration, Xtreme Bandwidth Architecture. XBA technology was invented during the design of our current 3D-graphics architecture. XBA consists of the eight-texel/four-pixel rendering pipeline, extremely wide 512-bit memory bus and our memory management logic. The memory management logic works as a highway system tying together the embedded DRAM memory, the external SDRAM memory, AGP memory and all units, which want access to the memory. A lot of the XBA inventions are behind this memory system, which handles 768 bits of data (eDRAM+AGP+SDRAM) every clock cycle, resulting in the massive bandwidth of 12.5 GB/s.
This huge bandwidth enables us to do everything in true color; we are not really interested in any 16-bit performance and dithered images. The bandwidth also allows us to do full-scene Anti-Aliasing with real supersampling. Although rendering game graphics at 1024x768 true-color resolutions requires around 2.5 GB/s of memory bandwidth, doing this with anti-alias enabled requires 10 GB/s of memory bandwidth! To say this in another way, products equipped with only external SDR or DDR memory are not able to do full-scene anti-aliasing at realistic speeds. The XBA architecture based products are really the first chips capable of improving the image quality with Anti-Aliasing, >>
Facebook
Twitter