Articles: Graphics
 

Bookmark and Share

(8) 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 ]

ATI Radeon HD 4850 X2 vs. Nvidia GeForce GTX 280: Face to Face

First, let’s compare the parameters of today’s fighters. The specifications and capabilities of the GeForce GTX 280 should be already well known to our readers, so we will begin with the newcomer. As we noted above, the ATI Radeon HD 4850 X2 is a 4850 CrossFire tandem designed on a single PCB just as the Radeon HD 4870 X2 represents the single-PCB version of a Radeon HD 4870 CrossFire configuration. This is reflected in its specs:

Click to enlarge

Click to enlarge

Indeed, these solutions only differ in terms of memory type and GPU clock rates, being identical in every other respect.

Comparing it with the GeForce GTX 280, the Radeon HD 4850 X2 has a tremendous advantage in computing resources, judging by the amount of ALUs. We should not forget about the superscalar design of the RV770’s shader processors, however: each of these 320 processors incorporates five ALUs, but not all of these five are utilized at every given moment. Moreover, Nvidia’s shader processors are clocked at a frequency of 1.3GHz as opposed to 625MHz of ATI’s solution. The DirectX 10.1 support is not a serious advantage, either. There are but few DirectX 10.1 applications available today, S.T.A.L.K.E.R.: Clear Sky being the only such game in our test programme.

The two cards seem to be equal in terms of texture processors, but Nvidia’s 80 texture processors often behave as only 40 processors effectively in real-life applications whereas the total amount of TMUs of the Radeon HD 4850 X2 is 80 if both GPUs are fully utilized by the application. Neither card seems to be superior in terms of rasterization processors whose efficiency depends on their internal design. Anyway, raster back-ends are not a decisive factor in modern games. More importantly, the Radeon HD 4850 X2, unlike its elder brother, has no advantage in terms of memory bandwidth. Its theoretical memory bandwidth is only 128GBps. The GeForce GTX 280 is somewhat better in this respect.

Thus, the current flagship of the GeForce series has good chances in the upcoming fight. Let’s now take a closer look at each fighter before they come to blows.

 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 ]

Discussion

Comments currently: 8
Discussion started: 01/12/09 04:23:12 PM
Latest comment: 09/09/09 08:44:42 PM

View comments

Add your Comment