Buy the ARP T-Shirt! BIOS Optimization Guide Money Savers!

 11 October 2006
 Discuss here !
Desktop Graphics Card Comparison Guide Rev. 33.0
Covering 628 desktop graphics cards, this comprehensive comparison allows you ... Read here
BIOS Option Of The Week - Virtualization Technology
Since 1999, we have been developing the BIOS Optimization Guide, affectionately known... Read here
Buy The BOG Book Subscribe To The BOG! Latest Money Savers!
Tech Myths B.U.S.T.E.D. Rev. 1.3
Digg! Reddit!Add to Reddit | Bookmark this article:

The More Pixel Pipelines A Graphics Card Has, The Faster It Is!

Yes, if we are back in the days when graphics dinosaurs ruled the Earth, two cards at the same clock speeds running a program that favours neither would generally net you the same performance. All else being equal, the number of pixel pipelines they have would be a fine indicator of their performance.

But as technology advances, the definition of a pipeline itself has gone down the drain, and this will be even more important as the future comes knocking at our door.

So, what makes up a traditional pipeline in those days? Generally, a simple pipeline of single functional units is like a conveyor belt with one worker doing one job at each step to produce a finished product at the end of the pipeline. With all kinds of different pipeline handling different jobs, it was pretty easy to differentiate performance, since one type of unit only did one type of job with a serial line of workers.

But as technology advances, so has the engineer's perspective of what we define as a pipeline. When ATI released their X1000 series of graphics cards, suddenly the definition of pipeline just fell apart. Why? Well, ATI decided to reinvent the pipeline.

Technically, an ATI Radeon X1950 graphics card has 16 pipelines, but each pipeline has three Shader Arithmetic Logic Units (ALUs), instead of the usual one per pipeline. That's why the Radeon X1950 is as fast as an NVIDIA GeForce 7950 GX2 (which has 48 pipelines), because when it comes to raw ALU power, it effectively matches the GeForce 7950 GX2's performance levels in games.

But that is just the start of it. As we know now, there are many pipelines with many units that do many different things inside a graphics card. For example, there are pixel pipelines, render output units (ROPs), texture units, etc. As we move closer to the future, the adoption of a unified architecture becomes more and more of a reality.

With a unified architecture, all units in a pipeline are general-purpose units. They do what they're told to do no matter what it is. To put it in layman terms, it's like having 24 workers who are good at everything, instead of 24 workers who are only good at one thing.

With Unified Architecture (UA), Graphics Processing Units (GPUs) will possess the ability to dynamically change the role of individual units based on need. For example, a shader-intensive game is going to need more shader processing power, rather than texturing power. With UA, the GPU can dynamically allocate more processing units to shader operations, instead of texture operations.

Isn't that a lot of silly fiddle-faddle as technology moves along? To avoid falling for marketing tactics, it's best to base our purchases of various performance metrics, like :

  • Game / application benchmarks
  • Synthetic benchmarks
  • Image quality

Those are the factors that we take into consideration in every review Tech ARP and othe tech websites produce. So, be sure to read our reviews and you will get the real deal!


Mobile CPU Comparison Guide Rev. 12.3
Western Digital AV-GP (WD30EURS) 3 TB Hard Disk Drive Review Rev. 2.0
The NVIDIA PhysX Tech Report
NVIDIA GeForce 9800 GTX+ Overclocking Guide
A Chat With Intel's Pat G
Zalman ZM600-HP 600W Power Supply Review
Seagate 160GB Momentus 5400.3 Hard Drive Review
Seagate 500GB Barracuda 7200.9 Hard Drive Review Rev. 2.0
OCZ PC2-5400 Gold DDR2 Memory Overview
Kingston 1GB PC2-5400 HyperX DDR2 SDRAM Dual Channel Memory Kit Review


Copyright © Tech All rights reserved.