 |
|||||||||||
|
|
|
||||||||||
   The Inno3D GeForce GT 220 scored 29% higher than the Radeon HD 4650 in 3Dmark Vantage at 1920x1200, giving it a clear performance advantage in this benchmark. However that said, the GeForce GT 220 was also 21% slower than the Radeon HD 4670, which delivered our best score of 1787pts at 1920x1200. |
||||||||||
|
||||||||||
|
Jason Saggers |
Posted on: 02/18/2010 02:26 AM
Wouldn't it make better sense to benchmark these cards with more budget CPU's, the CPU you are using would not be used by the end user that would be buying one of these budget cards. |
|
Steve Posts: 74 Joined: 2010-02-08 |
Posted on: 02/18/2010 05:30 AM
No not at all. When benchmarking GPU’s you want a fast CPU to remove any potential bottlenecks. Say in an extreme case we used a slow processor that created a system bottleneck, the GPUs would all be limited and possibly deliver the same performance. What would be the point of that? You will find this common practice amongst hardware sites. |
|
IFLATLINEI |
Posted on: 03/06/2010 09:23 PM
I understand why testers do this but it does nothing for those looking to see how these cards work with their pc's I dont see this card bottle necking much at all. So I think this is just a way to make it easier on your the tester or serving your own agenda rather than helping a consumer choose the right card for them. |
|
ProX |
Posted on: 03/06/2010 10:24 PM
IFLATLINEI seriously what are you talking about? Okay lets say they should use a slower processors, lets just for a moment assume that this makes sense (which it doesn't). What processor should they use? A $50 processor? Nah that's too cheap, well not for that guy but maybe for you. $100 processor? Yeah you can afford that so maybe a $150 processor then. Better yet lets use a processor that limits performance by 5 - 10fps and makes both graphics cards look the same. Yeah that's looking after the buyer, then they can spend more on a graphics card that is slower. I think what you guys are after is a budget buying guide which is a little different. That compares the value and performance of entire platforms and not just a single component. When testing components you cannot have any limitations. Its like testing two cars, a Nissan 370Z and a Ferrari 599 but limiting them to 30mph. In just about every test they would perform about the same so you could conclude that in terms of engineering the 370Z is better right? Take the limitations away and the Ferrari 599 would make the Nissan 370Z look silly. |
|
Steve Posts: 74 Joined: 2010-02-08 |
Posted on: 03/07/2010 01:31 AM
Posted by IFLATLINEI on 03/06/2010 10:23 PM
I can certainly understand how you might draw that conclusion but it is not the best way to test a graphics card. I like how you say “I dont see this card bottle necking much at all.” How much is not much at all and what experience are you basing this assumption on? Testing with the Core i7 processor is in no way easier for me, I have almost every processor currently being sold and there is no agenda here other than to display each graphics card in the best light. ProX is correct if you are after system articles please check the articles section where we have reviewed entire budget platforms. |