To test the HD 4670 IceQ Turbo 512Mb I will run a couple of artificial benchmark programs 3D Mark 06 and Vantage along with a couple of game benchmarks and compare the HD 4670 IceQ Turbo 512Mb with another graphics card of a similar value I have, the Nvidia 9600GTOC from BFG. The 3DMark programs will be run at default settings. Game settings will be shown under the respective results.Fitting the HD 4670 IceQ Turbo 512Mb was as easy as expected with no untoward issues.
I removed the drivers for my graphics card, powered down and replaced it with the HD 4670 IceQ Turbo 512Mb powered up and installed the latest drivers from the ATI website and rebooted and everything was working just fine and it was time to start testing. Once I had completed the testing of the HD 4670 IceQ I uninstalled it as I had uninstalled my original card and installed the 9600GT and reran all of the tests to get an idea of how the HD 4670 IceQ compared against a similarly priced card, that said the BFG 9600GTOC is approx £20 more expensive.
First I will run the games based benchmark programs.
The first game I chose was Counterstrike Source, why you might ask as it’s an old game, well many games run on the Source Engine so this should give a rough idea of how they would perform on the HIS HD 4670 IceQ.
Resolution was set to 1920x1200 and in “game” detail set to max.
HD 4670 IceQ Turbo CS:S results.
BFG 9600GTOC results.
I have to say I was most pleasantly surprised by these results from the HD 4670 IceQ Turbo, to achieve almost 150 frames per second through the in game benchmark at max settings at 1920x1200 resolution for a card costing around £70 is quite impressive. The 9600GTOC did a little better but then I fully expect it too as it is the more expensive of the 2 by some £20, which when you are talking £70 or £90+ is quite a chunk more cash.
The next game I chose was Farcry2 a newer game that even high end systems can find a bit challenging when the resolution and detail gets turned up.
The settings used were:-
Settings: Demo(Ranch Small), 1360x768 (60Hz), D3D10, Fixed Time Step(No), Disable Artificial Intelligence(No), Full Screen, Anti-Aliasing(None), VSync(No), Overall Quality(Optimal), Vegetation(Very High), Shading(Very High), Terrain(Very High), Geometry(Very High), Post FX(High), Texture(Very High), Shadow(Very High), Ambient(High), Hdr(Yes), Bloom(Yes), Fire(Very High), Physics(Very High), RealTrees(Very High)
HD 4670 IceQ Farcry2 results.
Min. Framerate: 38.04 / Average Framerate: 51.64 / Max. Framerate: 77.00
The 4670 struggled a little with the resolution turned up to 1920x1200, the native resolution of my monitor, but turning the resolution down a few notches and the anti-aliasing off provided the necessary relief to get playable frame rates again only occasionally dipping below the “golden” 30 frames per second.
Min. Framerate: 28.25 / Average Framerate: 41.93 / Max. Framerate: 60.50
The 9600GTOC fared somewhat better in Farcry2 at the same setting, as expected, never dropping below 30frame per second.
The final game benchmark I chose was Stalker Clear Sky, again this is a tough benchmark for even the most capable of systems. The resolution used was 1920x1200 default settings and no anti-aliasing with DX10.
HD 4670 Results.
The 4670 faired reasonably well at the settings I chose but could probably do with the resolution dropping down a bit to get the averages up to a “playable” 30 frames per second. For the purpose of the graphs I will use the lowest average, the sun shafts.
Again the 9600GTOC fared slightly better, the averages being a few frames per second better.
Artificial benchmarking programs results.
HD 4670 3D06 Result.
Unfortunately the HD4670 IceQ turbo didn’t fare anywhere near as well as the 9600GTOC in 3d06 with a deficit of 1700 and 1000 points in the Shader Model 2 and 3 scores respectively, maybe that £20 price difference is beginning to really show.
HD4670 IceQ Turbo Vantage results.
9600GTOC Vantage Result, again we see quite a significant difference between the two cards with approaching a 50% increase in frames per second on the 2 GPU tests for the more expensive 9600GTOC.
I did expect the 9600GTOC to be slightly better after the results from the in game benchmarks, but not quite that much. It just goes to show, artificial benchmarks are maybe a little too artificial as “in game” there is nothing like that kind of difference between the two cards.
Next Page - Conclusion