Thursday 15 November 2012

Benchmark System

We first benchmarked all the graphics cards in our VGA 2012 benchmark system. After that, we took a look at installation and operation in a closed case.

Benchmark System
CPUIntel Core i7-2600K (Sandy Bridge), 32 nm, 4C/8T, 8 MB Shared L3 Cache, Hyper-Threading Enabled, Overclocked to 4.5 GHz
MotherboardGigabyte Z68X-UD7-B3, Intel Z68 Express, BIOS F10
Memory4 x 4 GB Kingston HyperX DDR3-1600
CPU CoolerCooler Master TPC 800 with Noiseblocker eLoop B12-PS (PWM)
SSDKingston V200+ 480 GB
Power SupplyCorsair AX1200i Digital ATX Power Supply, 1200 W, 80 PLUS Platinum
Operating SystemWindows 7 x64 Ultimate
Driver Catalyst 12.11 (Beta) and Catalyst 12.9
GeForce 306.97 WHQL
Temperature22 °C (held constant)
Closed Benchmark CaseNZXT Phantom 820

We're particularly interested to compare the performance of AMD's Catalyst 12.9 driver package to the more recently-released 12.11 beta driver, which the company claims should be much faster in a number of games.

Zoom


Zoom

Unlike many of the graphics card reviews you read, we're conducting our testing today in a closed chassis. After all, it can't be that easy to dissipate more than 500 W worth of heat, and we want to know what effect that has on real-world thermals and acoustics.Benchmark Results: Synthetics

Benchmark Results: Synthetics

When you see some of these cards show up twice, remember that we're testing the AMD board with two driver versions. The Catalyst 12.11 beta package propels performance forward quite a bit (in some cases well beyond Nvidia's results). We didn’t think that new drivers would make this much of a difference (again), but AMD manages to give its cards a very welcome boost. We're looking into what the company did to achieve its speed-up.

Unigine Heaven 2.5

Clearly, AMD put a lot of effort into optimizing the Catalyst 12.11 driver for synthetic benchmark performance. But do actual games benefit just as much from the company's tuning?
We didn’t bother with Catalyst 12.9 in Unigine, but we did include a triple-monitor setup.

Unigine Sanctuary

We’re only using the Catalyst 12.11 beta build in this benchmark as well. The average frame rate increase moving from Catalyst 12.9 to 12.11 is only about five percent, anyway.

No comments:

Post a Comment