Source: Tom's hardware UK – Keywords: HDTV, Nvidia, ATI
Categories: Graphics
Test results : CPU occupation
The average CPU usage rate on our two platforms was as follows:
Long live HD video engines
Firstly, we noticed that cards not equipped with Purevideo 2 and Avivo HD (8800 GTS and HD 2900 XT here) engines were not sufficient to read HD H.264 films on an old Athlon 64. On the other hand, the Radeon HD 2400 and 2600 and the Geforce 8600 will let you watch a H.264 film on this processor quite comfortably.
The good news from this test is that the purchase of a card equipped with a HD video engine can rejuvenate an outdated setup. Standing shoulder to shoulder with a Radeon HD 2400 or a 2600, our Athlon 64 3800+ does not go over 35% of use. Because of this, we enquired the minimal CPU frequency at which our HD films could be read at. Our tests showed that this was 1600 MHz.
This goes to show that the frequency of an Athlon is 64 2800+ in socket 939 or 3000+ in socket 754. This is a very pleasant surprise in that it opens up many interesting perspectives for the manufacture of Cheap and quiet HD boxes.
While we didn’t test discs with an average bit rate of 25Mb/s the formats go up to 30 Mb/s VC1 for HD DVD and 40 Mb/s for Blu-ray so 20 and 60% more respectively. The Radeon HD 2900 XT and Geforce 8800 then begin to show their limits. Finally, if you prefer Nvidia over ATI, know that you will then need to go up a notch in frequency, 2 GHz is the minimum to comfortably read a film in VC-1 with a Geforce 8600 GT.
Long live dual-core processors
The first major conclusion of this test is that any graphics card, whether it includes a HD video engine or not, are sufficient to read HD films given that the processor is a dual-core. Keep in mind though that this conclusion isn’t valid for films with an average rate of more than 25 Mb/s.
The ATI advantage
If we compare the effectiveness of Nvidia and ATI engines, we see that the latter has a net advantage over its competition for the reading of films in VC-1, which is perfectly in line with the manufacturer’s specifications. The Purevideo 2 seems to be identical to the Purevideo 1 when it comes to the decoding of the VC-1: the Geforce 8600 and the 8800 have the same performance. On the other hand for the reading of H.264, the gap is much smaller and constantly fluctuating. On the Athlon, the Geforce seems more effective but it falls behind the Radeon HD 2600 XT on the Core 2 Duo. The Radeon HD 2400 XT also showed surprising results. While it seems to be twice as powerful as its predecessor HD 2600 XT on the Athlon, it proves around half as powerful once mounted with a Core 2 Duo. This is surprising, as the two cards are normally equipped with the same engine. A possible slant in our test could have contributed to the fact that the HD 2400 XT ran on the catalyst 7.6 beta because of its incompatibility with the 7.7 WHQL… However we have not noted any performance gap between these drivers and the latest Catalyst 7.7 WHQL with the Radeon 2600 XT.
- Previous page Tests : materials, methods and problems
- Next page Test results : image quality
- How Cool Are Thermalright's Graphics Card Coolers?
- The Best Gaming Video Cards for the Money UK: August 2007
- Leadtek's GeForce 8800 Performance Leviathan
- HD 2600 & GeForce 8600: Where's the Mid-Range?
- San Base: Computer Graphics Avant-Garde
- The Best Gaming Video Cards for the Money: July 2007
- The Best Gaming Video Cards for the Money: June 2007
- The Best Gaming Video Cards for the Money: May 2007
- Foxconn's GeForce 8600 GTS
- Spring VGA Charts 2007
I think it would be a really good idea to get this article edited by a native English speaker.
Other than that, very timely as am at the start of the upgrade cycle and find myself having to buy a PCI-E graphics card.
To be fair, the Radeon 2400 should have been compared against a 8400GS, as they are the same price, and any question about drivers affecting performance could have been answered by running the 2600 on the same drver as the 2400.