Page 1:Finding Value In Higher Quality
Page 2:CPU, Cooler, And Memory
Page 3:Graphics, Motherboard, And Power Supply
Page 4:Case And Drives
Page 5:Assembling Our $2000 Performance PC
Page 6:Getting Our Core i7-3930K To 4.6 GHz
Page 7:Pushing GeForce GTX 670 To Its Limit
Page 8:Test Settings And Benchmarks
Page 9:Benchmark Results: 3DMark And PCMark
Page 10:Benchmark Results: SiSoftware Sandra
Page 11:Benchmark Results: Battlefield 3 And DiRT 3
Page 12:Benchmark Results: Skyrim And StarCraft II
Page 13:Benchmark Results: Audio And Video Encoding
Page 14:Benchmark Results: Productivity
Page 15:Power, Heat, And Efficiency
Page 16:Sometimes, Lower Value Is OK
Power, Heat, And Efficiency
Our previous build has three big advantages when it comes to power consumption: first, its 22 nm Ivy Bridge-based CPU; second, its quad-core design is less complex than our Sandy Bridge-E-based solution; and third, an 80 PLUS Gold-rated power supply is more efficient than today's Silver-rated alternative.
On the other hand, our newer system should save a little bit of power by virtue of its GeForce GTX 670 graphics card, which doesn't consume as much as the GeForce GTX 680.
The difference in CPU load power is stark, especially when our new machine is overclocked. We had hoped that its 850 W power supply would be sufficient for three-way SLI arrays, but that option would only be completely safe at a more conservative overclock.
Our attempts to get more air across our CPU heat sink in the NZXT Phantom 410 (without spending money on additional fans) were not able to overcome the chip's higher thermal output or the cooler's lower performance. As a result, its temperatures are barely within the bounds of what we'd call acceptable.
This quarter's build outperforms last quarter's, even in the average of game tests (a result of the lower resolutions used for comparison against the other two builders' machines). Don't worry, though. We’ll discuss our results at 2560x1600 in the conclusion.
Each of our application suites (games, encoding, productivity) make up 30% of our average performance number, while real-world-based synthetic storage tests make up the other 10%. The better performance of today's build, unfortunately, is not enough to offset its higher power consumption, and its comparative efficiency is down by up to 28%!
- Finding Value In Higher Quality
- CPU, Cooler, And Memory
- Graphics, Motherboard, And Power Supply
- Case And Drives
- Assembling Our $2000 Performance PC
- Getting Our Core i7-3930K To 4.6 GHz
- Pushing GeForce GTX 670 To Its Limit
- Test Settings And Benchmarks
- Benchmark Results: 3DMark And PCMark
- Benchmark Results: SiSoftware Sandra
- Benchmark Results: Battlefield 3 And DiRT 3
- Benchmark Results: Skyrim And StarCraft II
- Benchmark Results: Audio And Video Encoding
- Benchmark Results: Productivity
- Power, Heat, And Efficiency
- Sometimes, Lower Value Is OK