Performance and Conclusions

Since the point of any overclocking is to improve performance, we tested it with 3DMark11 benchmark (v 1.01, released just before posting of this article) and canned stand-alone benchmarks for Aliens vs. Predator and Lost Planet 2. The system was perfectly stable through all of this testing, so further overclocking is possible.

3D Performance
Test
Settings
Stock
OC
3DMark11

(v 1.01)
Performance
3878
4029
Extreme
1360
1494
Test
Settings
Avg. FPS
Avg. FPS
Alien vs. Predator Benchmark
1920x1200, maxed except 2xAA, 8xAF, advanced shadows disabled
36.0
39.2
Lost Planet 2 DX9 Benchmark
1920x1200, maxed
59.5
59.9
Lost Planet 2 DX11 Benchmark
1920x1200, maxed except 4xAA, medium shadow/texture detail, medium DX11 features
27.4
36.0

Nice gains were seen in some benchmarks, but as usual, the benefits varied from game to game. We expect that if a dozen modern games were benchmarked, the improvements would be in the range of 7~20%. A quick trial on a couple of the benchmarks with only the video card overclocked suggests that a greater portion of the gains are due to the overclock of the GPU.

Interestingly, even when the various benchmarks were run back-to-back with nary a pause, up to a total of over 15 runs in 2-3 hours, much of it with both GPU and CPU overclocked, the system stayed very quiet. There was no increase in noise beyond the 18 dBA@1m maximum measured under extended stress testing with Furmark and CPUBurn. The PC never reached even 300W in any of the benchmark testing.

Note that some of the same benchmarks were used at the same settings to test our SFF Silent Gaming PC. The results of the two systems were very close, despite the SFF box using an Intel i5 processor; they both used the same HD 6870 video card.

3D Game FPS: SFF Gamer vs. Mid Gamer
Test
Stock Mid Gamer
OC Mid Gamer
SFF Gamer
Alien vs. Predator Benchmark
36.0
39.2
37.9
Lost Planet 2 DX9 Benchmark
59.5
59.9
58.5
Lost Planet 2 DX11 Benchmark
27.4
36.0
36.9

CONCLUSIONS

The aim of this article was to show how to assemble a silent gaming PC of mid-level performance and price. The maximum noise levels under extreme lab test conditions in its overclocked state was a bit louder than we would have liked, but most likely, it is still 10~15 dBA@1m quieter than similar performance gaming rigs espoused by other tech sites or sold by most system integrators. During typical 3D game play, it is difficult to push the acoustics of this rig up beyond 20 dBA@1m, overclocked or not. And measuring just 14 dBA@1m at idle, it challenges the noise levels of even fanless PCs.

We will be flattered to find posts in our forums (and elsewhere) of readers who follow our guide A-to-Z but it would be even more gratifying to read about an alternate DIY system inspired by our guide. The point is not necessarily to duplicate SPCR's Mid Gamer Silent Rig, but to consider our approach and apply it to your own chosen components. It is feasible and fun to build powerful gaming rigs that don't resemble jet planes at takeoff.

* * *

Other Silent PC Build Guides:
SFF Silent Gaming PC
HTPC Home Server
Gaming/Home Server: HD 5780 + 9 HDDs
Mid-Tower Home Server Configurations
SFF Home Server Configurations

* * *

Discuss this article in the SPCR Forums.