Graphics Card Comparison
Click here for Graphics Card
Comparison 2 includes Geforce 6 Series and Radeon X Series
Click here for Graphics Card Comparison
3 includes Geforce 7 Series and Radeon X1000 series
Graphics Cards are one of the most upgraded features in a PC, especially if you are a PC gamer. Picking the next card for your machine could be a problem if you need specific features from it. This page was written to show you the differences and benefits of the latest graphics cards in comparison with the others on the market.
You can compare the clock speeds, memory speeds and types as well as polygon and pixel counts. Then on to Memory bandwidth and the technology features that make each card special in its own way. Not forgetting of course you can compare the prices of all these cards as well, which is the most important thing for most of us.
| GFX Card | Archit-ecture | Clock Speed | Memory Speed | Memory Size | Memory Bandwidth | RAM-DAC | Fill Rate Pixels/s | Price |
| Geforce 2 MX | 256-bit | 175Mhz | 166Mhz | 32Mb | 2.7Gb/s | 350Mhz | 350 Million/s | Compare |
| Geforce 4 MX 420 | 256-bit | 250Mhz | 166Mhz (SDR) | 64Mb | 2.7Gb/s | 350Mhz | 500 Million | Compare |
| Geforce 4 MX 440 | 256-bit | 270Mhz | 400Mhz (DDR) | 64Mb | 6.4Gb/s | 350Mhz | 540 Million | |
| Geforce 4 MX 460 | 256-bit | 300Mhz | 550Mhz (DDR) | 64Mb | 8.8Gb/s | 350Mhz | 600 Million | |
| Geforce 4 Ti 4200 | 256-bit | 250Mhz | 500Mhz (DDR) |
64Mb /128Mb |
8.0Gb/s | Dual 350Mhz | 1 Billion | |
| Geforce 4 Ti 4400 | 256-bit | 270Mhz | 550Mhz (DDR) | 128Mb | 8.7Gb/s | Dual 350Mhz | 1.1 Billion | |
| Geforce 4 Ti 4600 | 256-bit | 300Mhz | 600Mhz (DDR) | 128Mb | 10.4Gb/s | Dual 350Mhz | 1.2 Billion | |
| Geforce 4 Ti 4800 | 256-bit | 300Mhz | 650Mhz (DDR) | 128Mb | 10.4Gb/s | Dual 350Mhz | 1.2 Billion | |
| Geforce 4 Ti 4800SE | 256-bit | 275Mhz | 550Mhz (DDR) | 128Mb | 8.8Gb/s | Dual 350Mhz | 1.1 Billion | |
| Geforce FX 5200 | 256-bit | 250Mhz | 400Mhz (DDR) | 128Mb | 6.4Gb/s | Dual 400Mhz | 1 Billion | Compare |
| Geforce FX 5200 Ultra | 256-bit | 350Mhz | 650Mhz (DDR) | 128Mb | 10.4Gb/s | Dual 400Mhz | 1.4 Billion | |
| Geforce FX 5600 | 256-bit | 325Mhz | 550Mhz (DDR) | 256Mb | 8.8Gb/s | Dual 400Mhz | 1.3 Billion | Compare |
| Geforce FX 5600 Ultra | 256-bit | 350Mhz | 800Mhz (DDR) | 256Mb | 12.8Gb/s | Dual 400Mhz | 1.4 Billion | |
| Geforce FX 5700 Ultra | 256-bit | 475Mhz | 900Mhz (DDR) | 256Mb | 14.5Gb/s | Dual 400Mhz | 1.9 Billion | Compare |
| Geforce FX 5800 | 256-bit | 400Mhz | 800Mhz (DDR) | 128Mb | 12.8Gb/s | Dual 400Mhz | 1.6 Billion | Compare |
| Geforce FX 5800 Ultra | 256-bit | 500Mhz | 1Ghz (DDR) | 128Mb | 16Gb/s | Dual 400Mhz | 4.0 Billion | |
| Geforce FX 5900 Ultra | 256-bit | 450Mhz | 850Mhz (DDR) | 256Mb | 27.2Gb/s | Dual 400Mhz | 1.8 Billion | Compare |
| Geforce FX 5950 | 256-bit | 475Mhz | 950Mhz (DDR) | 256Mb | 30.4Gb/s | Dual 400Mhz | 1.9 Billion | Compare |
| Matrox Parhelia | 512-bit | 220Mhz | 550Mhz (DDR) | 128Mb | 17.6Gb/s | Dual 400Mhz | 880 Million | Compare |
| Radeon 7500 | 256-bit | 290Mhz | 460Mhz (DDR) | 64Mb | 7.4Gb/s | 350Mhz | 580 Million | Compare |
| Radeon 8500 | 256-bit | 275Mhz | 550Mhz (DDR) | 64Mb | 8.8Gb/s | Dual 400Mhz | 1.1 Billion | Compare |
| Radeon 9000 Pro | 256-bit | 275Mhz | 550Mhz (DDR) | 64Mb | 8.8Gb/s | Dual 400Mhz | 1.1 Billion | Compare |
| Radeon 9500 Pro | 256-bit | 275Mhz | 550Mhz (DDR) | 128Mb | 8.6Gb/s | Dual 400Mhz | 2.2 Billion | Compare |
| Radeon 9600 Pro | 256-bit | 400Mhz | 600Mhz (DDR) | 128Mb | 9.6Gb/s | Dual 400Mhz | 1.6 Billion | Compare |
| Radeon 9700 Pro | 256-bit | 325Mhz | 620Mhz (DDR) | 128Mb | 19.4Gb/s | Dual 400Mhz | 2.6 Billion | Compare |
| Radeon 9800 Pro | 256-bit | 375Mhz | 675Mhz (DDR) | 128Mb / 256Mb | 21.8Gb/s | Dual 400Mhz | 3.04 Billion | Compare |
| Radeon 9800 XT | 256-bit | 412Mhz | 730Mhz (DDR) | 256Mb | 23.4Gb/s | Dual 400Mhz | 3.296 Billion |
Remember if your looking for newer graphics cards head over to graphics card comparison 2 and graphics card comparison 3
Its worth noting that the figures above are based on the raw power of the cards listed. However power is nothing without direction and control. because a graphics card has more Megahertz than another is a good sign that the card is faster but its not always the case. Technology behind these chipsets provides a more in depth look at which card we should purchase.
Examples of this are with pixel pipelines. Cards with more pixel pipelines. When a pixel is processed it passes through the (or one of) the pixel pipeline. Newer cards have a greater amount of pixel pipelines and so can render more pixels per cycle than ones with fewer pixel pipelines.
For example we have card A and card B, Card A has 2 pixel pipelines and a Core clock of 500Mhz, card B has 4 pixel pipelines and has a core clock of 300Mhz., If you look and many specs on manufacturers sites etc they will state that there card has 500Mhz core clock. You look at that then at card B with a 300Mhz clock and presume that card A is faster. In Terms of pixels/s that is incorrect. Take a look at the calculation
Card A 500Mhz (500Million cycles/s) * 2 Pixel Pipelines = 1 Billion pixels rendered per second
Card B 300Mhz (300Million cycles/s) * 4 Pixel Pipelines = 1.2 Billion pixels rendered per second.
There is far more than this going on with graphics cards but the example above is designed to show you that sheer power is not enough to become the best graphics card.
The table below will show you some of the technology features of the graphics cards to give you a more in depth look at there differences.
| GFX Card | Pixel Pipelines | Texture units / pipeline | Textures per clock | Memory Bus Width | New Technologies brought into the card |
| Geforce 2 MX | 2 | 2 | 4 | 128-Bit | T&L (second generation) |
| Geforce 4 MX 420 | 2 | 4 | 8 | 128-Bit |
Lightspeed Memory - Architecture II. NView Display. Accuview II |
| Geforce 4 MX 440 | 2 | 4 | 8 | 128-Bit | |
| Geforce 4 MX 460 | 2 | 4 | 8 | 128-Bit | |
| Geforce 4 Ti 4200 | 4 | 2 | 8 | 128-Bit |
NfiniteFX II engine. Lightspeed Memory - Architecture II. NView Display. Accuview II Dual Vertex Shaders |
| Geforce 4 Ti 4400 | 4 | 2 | 8 | 128-Bit | |
| Geforce 4 Ti 4600 | 4 | 2 | 8 | 128-Bit | |
| Geforce 4 Ti 4800 | 4 | 2 | 8 | 128-Bit | |
| Geforce 4 Ti 4800SE | 4 | 2 | 8 | 128-Bit | |
| Geforce FX 5200 | 4 | 2 | 8 | 128-Bit |
CineFX DDR-II (5800 only) 8x AGP NView Display Intellisample Technology |
| Geforce FX 5200 Ultra | 4 | 2 | 8 | 128-Bit | |
| Geforce FX 5600 | 4 | 2 | 8 | 128-Bit | |
| Geforce FX 5600 Ultra | 4 | 2 | 8 | 128-Bit | |
| Geforce FX 5800 | 4 | 2 | 8 | 128-Bit | |
| Geforce FX 5800 Ultra | 4 | 2 | 8 | 128-Bit | |
| Geforce FX 5900 Ultra | 4 | 2 | 8 | 256-Bit |
CineFX 2.0 DDR-II 8x AGP NView Display Intellisample |
| Geforce FX 5950 Ultra | 4 | 2 | 8 | 256-Bit | |
| Geforce 6800 Ultra | 16 | 1 | 16 | 256-bit |
CineFX 3.0 GDDR-3 memory Intellisample 3.0 High-Precision Dynamic-Range (HPDR) Technology UltraShadow II |
| Matrox Parhelia | 4 | 4 | 16 | 256-Bit |
16x FSAA TV Encoder Triple monitor support |
| Radeon 7500 | 2 | 3 | 16 | 128-Bit |
FSAA Hyper Z Charisma Engine |
| Radeon 8500 | 4 | 2 | 8 | 128-Bit |
Truform Smoothvision Hyper Z II Charisma Engine II Pixel Tapestry II |
| Radeon 9000 Pro | 4 | 1 | 4 | 128-Bit | |
| Radeon 9500 Pro | 8 | 1 | 8 | 128-Bit |
Hyper Z III Smoothvision 2.0 SmartShader 2.0 Truform 2.0 Video Shader |
| Radeon 9600 Pro | 4 | 1 | 4 | 128-Bit |
Smoothvision 2.1 SmartShader 2.1 Video Shader Hyper Z III |
| Radeon 9700 Pro | 8 | 1 | 8 | 256-Bit |
Hyper Z III Smoothvision 2.0 SmartShader 2.0 Truform 2.0 Video Shader |
| Radeon 9800 Pro | 8 | 1 | 8 | 256-Bit |
Smart Shader 2.1 Hyper Z III+ Smoothvision 2.1 |
| Radeon 9800 XT |
As you can see graphics cards vary wildly in terms of what they have to offer. However all the new technologies are normally around the same when it comes to performance from cards from the same time. The best Nvidia card is neither to far ahead or behind the top ATI offering. Preference plays a big part in deciding which card you require.
Improvements are being made all the time, and as you can see from the lists not always in the form of speed. This is how the graphics card marker differs from the CPU market. CPU haven't managed to shrug off the idea that faster is always better. This is why we have AMD using PR numbers instead of clock speed. The graphics card industry realises there are other ways to make improvements other than raw power.
Should any of these specifications be incorrect please do no hesitate to contact me at webmaster@pantherproducts.co.uk with the correct info I will be only to happy to alter it