View Single Post
Old 02-01-11, 11:34 PM   #5
ViN86
 
Join Date: Mar 2004
Posts: 15,486
Default Re: nV News GTX 560 Ti Review - EVGA SuperClocked

Results: Games

Let’s begin with the game benchmarks. All benchmarks listed here utilized in-game benchmarks when available, or FRAPS benchmark for games where an in-game tool was not available. All benchmarks have been repeated multiple times to check for agreement, and the values were then averaged together. Furthermore, all tests here used a resolution of 1920x1200.



Metro 2033

First up is Metro 2033. This game released in March 2010 features DirectX 11 and is known to tax some of the latest video cards. The benchmarking tool was used at two different settings to distinguish the card’s ability to handle moderate levels of AA/AF and maximum values. The tests conducted are listed below.
  • Test A: Medium Quality, AAA, 4xAF, PhysX On, DOF Off, Tessellation On
  • Test B: Very High Quality, 4xMSAA, 16xAF, PhysX On, DOF On, Tessellation On

Test A represents moderate settings utilized by most gamers and Test B represents the highest quality graphics (in the benchmarking tool options). The results are shown in the graph below.



Here we can see the GTX 560 Ti is off to a good start. At medium settings, the card easily bests the GTX470 and the HD5870 and offers playable framerates for gamers. At the highest settings, the card is on par with the HD5870, but performs slightly worse than the GTX470. The HD5870 can’t seem to keep up with the GTX470 and GTX 560 Ti in Tessellation tests (as will be seen), but at higher levels of AA/AF, the 470 pulls ahead. While none of the cards offer playable framerates at these high settings, the test is a good stress test for all the cards. Next up is Resident Evil 5.



Resident Evil 5

Resident Evil 5 was released for PC in September 2009 and features DirectX 10 graphics. This test was added to show the GTX 560 Ti’s performance in DirectX 10 games. Once again, two different tests will be used with different settings. The first test will use medium settings with no AA and the second test will use the highest possible in-game settings and 8xAA. The in-game Fixed Benchmark was used for the following tests.
  • Test A: Medium Settings, No AA, Motion Blur On
  • Test B: High Settings, 8xAA, Motion Blur On



The GTX 560 Ti performed very well in both tests. At medium settings, the GTX 560 Ti performed on par with the GTX470, and both cards beat out the HD5870. At the highest settings, the three cards performed well and all offer playable frame rates. The GTX 560 Ti performed especially well, beating the other two cards by 10 FPS. Next up is H.A.W.X. 2.



H.A.W.X. 2

H.A.W.X. 2 was released in November 2010 and features DirectX 11. The game uses tessellation for its terrain rendering and features amazing aerial views. Two settings were used for the tests to compare the cards’ abilities to handle Tessellation and AA. Test A used highest settings with No AA and the Terrain Tessellation Off, and Test B used the highest settings with 8xAA and Terrain Tessellation On.
  • Test A: Highest Settings, No AA, Tessellation off
  • Test B: Highest Settings, 8xAA, Tessellation On



Here we can see the GTX 560 Ti shine. The card easily outperforms the HD5870 and beats the GTX470 by over 10 FPS in both cases. All three cards offer playable frame rates at both settings, but the GTX 560 Ti offers the best performance by far. Next up is Call of Duty: Black Ops. This test will showcase the card’s DirectX 9 performance.



Call of Duty: Black Ops

Call of Duty: Black Ops was released in November 2010 and uses DirectX 9. The game is the latest installment of Activision’s Call of Duty franchise. Two tests were performed with one at normal settings and no AA/AF and the other at maximum in-game settings. Since no in-game benchmark utility exists, demos were recorded in-game and tested using the timedemo feature. The demo featured five minutes of game play on Havana. The tests performed are as follows.
  • Test A: No AA, No AF, Normal Texture Quality, Bilinear Texture Filtering
  • Test B: 8xAA, 16xAF, Extra Texture Quality, Trilinear Texture Filtering



In Black Ops, all three cards performed very well. There wasn’t much difference between the frame rates with no AA/AF and those with 8xAA/16xAF. This is somewhat expected since modern cards are very adept at handling DirectX 9 games, however, the difference between the two frame rates seems to be too small. There may be some problems with CPU limiting as well. The game has been known to have CPU optimization issues and this may be the cause of the relatively low frame rates at low settings. While the GTX 560 Ti lags slightly behind the HD5870 and GTX470, the frame rates are well above the minimum playable frame rate. The fifth and final in-game test performed was Battlefield: Bad Company 2.



Battlefield: Bad Company 2

Battlefield: Bad Company 2 was released in March 2010 and features DirectX 11. Since no in-game demo exists, the first mission in the campaign was tested, from the end of the opening cinematics to the beginning of the first cutscene and FRAPS was used to record frame rates. Two settings were used for the tests, including on test at High settings with 8xMSAA/16xAF and HBAO On, and the other test with Medium settings, 4xMSAA/4xAF and HBAO Off.
  • Test A: Medium Settings, 4xMSAA, 4xAF, HBAO Off
  • Test B: Highest Settings, 8xMSAA, 16xAF, HBAO On



The GTX 560 Ti performs solidly at both settings and beats the GTX 470 and HD5870 at medium settings. At high settings, the GTX 560 Ti matches the performance of the HD5870. Both the GTX 560 Ti and the HD5870 beat the GTX 470 at high settings. It seems like the GTX 560 Ti easily beats the competition in low AA/AF situations, but at higher levels of AA/AF, the larger number of shaders on the HD5870 and GTX 470 give the two cards a slight advantage. However, the GTX 560 Ti makes up for this shortcoming with raw speed. If this is true, then we should expect to see the GTX 560 Ti pull ahead when overclocked.

Next up are the synthetic benchmarks. The tests performed included 3DMark Vantage, and the "new standard trifecta" of 3DMark 11, Unigine’s Heaven 2.1 Benchmark, Bitsquid’s Stone Giant demo, which all test DirectX 11 capabilities such as hardware tessellation.

Last edited by ViN86; 02-03-11 at 05:48 PM. Reason: update
ViN86 is offline