Settings and methods of testing
All graphics cards joining in the test were
tested in a system with the following configuration:
·
Mainboard: Intel Siler DX79SI (Intel
X79 Express, LGA 2011, BIOS 0494 from 7/23/2012)
·
CPU: Intel Core i7-3960X Extreme Edition,
3.3GHz, 1.2 V, 6x256KB L2, 15MB L3 (Sandy Bridge-E, C1, 32nm)
·
CPU cooler: Phanteks
PH-TC14PE (2x135mm fans at 900RPM)
·
Thermal interface: ARCTIC MX-4
·
Graphics cards: Asus GeForce GTX 660 Ti DirectCU
II TOP 2GB (TI-DC2T-2GD5); Asus GeForce GTX 660 DirectCU II OC 2GB
(GTX660-DC2O-2GD5); Sapphire Redeon HD 7870 GHz Edition OC 2GB; HIS 7850 IceQ
Turbo X 2GB (H785QT2G2M)
·
System memory: DDR3 4x 4GB Mushkin
Redline (Spec: 2133MHz / 9-11-10-28 / 1.65V)
·
System hard drive: Crucial m4 256GB
SSD (SATA-III,CT256M4SSD2, BIOS v0009)
·
Drive for programs and games: Western Digital
VelociRaptor (300GB, SATA-II, 10000RPM, 16MB cache, NCQ) inside Scythe Quiet
Drive 3.5” HDD silencer and cooler
·
Backup drive: Samsung Ecogreen F4
HD204UI (SATA-II, 2TB, 5400RPM, 32MB, NCQ)
·
Case: Antec Twelve Hundred (front
panel: three Noiseblocker NB-Multi-frame S-Series MF12-S2 fans at
1020RPM; back panel: two Noiseblocker NB-BlackSilentPRO PL-1 fans at
1020RPM; top panel: standard 200mm fan at 400RPM)
·
Control and monitoring panel: Zalman ZM-MFC3
·
Power supply: Seasonic SS-1000XP Active PFC F3
1000 W (with a default 120mm fan)
·
Monitor: 27” Samsung S27A850D (DVI-I, 2,560x1,440,
60Hz)
GeForce GTX 660 Graphics Cards
Besides the graphics
cards discussed above we also added five more products to the performance
comparison charts. Firstly, the source Nvidia GeForce GTX 660 2GB works at
980/6008MHz frequencies which will help us judge the advantages of the factory
overclocked products against the standard solution. Secondly, Asus GeForce
GTX 660 Ti DirectCU II TOP 2 GB (TI-DC2T-2GD5) is included for comparison
against one of the overclocked GeForce GTX 660 graphics cards from today’s test.
Thirdly, The Asus GeForce GTX 680 DirectCU II TOP 2 GB (GTX680-DC2T-2GD5) was
added for comparison against a pair of GeForce GTX 660 based graphics cards which
is tied with each other in a 2-way SLI setting. Finally, we also added two
graphics cards using AMD processors, named Sapphire Radeon HD 7870 GHz
Edition OC 2 GB and Sapphire Radeon HD 7970 OC Dual-X 3 GB. I believe
there’s no necessary comment to explain why we chose these contestants.
Moreover, I would like to add that like all other testing contestants, these
additional graphics cards were also tested at their initial clock speeds (while
the factory overclocking is right from the start).
As usual, so as to reduce the dependence of
the graphics cards performance on the overall platform speed, I overclocked our
32nm six-core CPU with the BCLK frequency set at 125MHz and “Load-Line
Calibration” allowed at 4.625GHz. The processor’s Vcore was raised up to 1.49V
in the mainboard BIOS.
CPU-overclocking
Hyper-Threading technology was activated.
16GB of system DDR3 memory worked at 2GHz frequency with 9-10-10-28 timings plus
1.65V voltage.
The test was carried out in Microsoft
Windows 7 Ultimate x64 SP1 with all updates and drivers, at that day, as
following:
·
Intel Chipset Drivers 9.3.1.1009 WHQL from 12/08/2012 for
the mainboard chipset
·
DirectX End-User Runtimes libraries
from November 30, 2010
·
AMD Catalyst 12.11 beta (9.010.8) driver
from 12/02/2012 + Catalyst Application Profiles 12.11 (CAP2) for AMD based
graphics cards
·
Nvidia GeForce 310.70 beta driver from
12/04/2012 for Nvidia based graphics cards
As we have
already tested GeForce GTX 660 generally in one of our previous
reviews, today we will check out only one resolution – 1,920x1,080 pixels. The
tests were done in two image qualities: “Quality+AF16x” – default texturing
quality in the drivers with 16x anisotropic filtering enabled and “Quality+
AF16x+MSAA 4(8)x” with 16x anisotropic filtering enabled and full screen 4x or
8x antialiasing if the average frame rate was high enough for enjoyable gaming
experience. We also activated anisotropic filtering and visual anti-aliasing while
adjusting the game settings. If these corresponding options were not enough, we
changed them in the Control Panels > Catalyst and GeForce driver sections.
We also disabled Vsync. There were no further changes in the driver settings.
The list of games and applications that
were in this test session was shortened and included two benchmarking standards
and 8 newest games featuring various genres with all updates installed by the
time of the test taken:
Games participating in the test:
·
3DMark 2011 (DirectX 11) – version 1.0.3.0,
Performance and Extreme profiles;
·
Metro 2033: The Last Refuge (DirectX 10/11)
- version 1.2, maximum graphics quality settings, official benchmark, “High”
image quality settings; DOF and MSAA4x disabled; AAA aliasing enabled, two
consecutive operations
·
Total War: Shogun 2: Fall of the Samurai (DirectX
11) – version 1.1.0, built-in benchmark (Sekigahara battle) at maximum graphics
quality settings and enabled MSAA 4x in one of the test modes
·
Crysis 2 (DirectX 11) – version 1.9, we
used Adrenaline Crysis 2 Benchmark Tool v.1.0.1.14. BETA with “Ultra
High” graphics quality profile and activated HD textures, two operations of a
demo recorded on “Times Square” level;
·
Battlefield 3 (DirectX 11) – version 1.4,
all image quality settings set to “Ultra”
·
Sniper Elite V2 Benchmark (DirectX 11) –
version 1.05, using Adrenaline Sniper Elite V2 Benchmark Tool v1.0.0.2
BETA with maximum graphics quality settings (“Ultra” profile), Advanced
Shadows: HIGH, Ambient Occlusion: ON, Stereo 3D: OFF
·
Sleeping Dogs (DirectX 11) – version 1.5, using Adrenaline
Sleeping Dogs Benchmark Tool v1.0.0.3 BETA with maximum image quality
settings, Hi-Res Textures pack installed, FPS Limiter and V-Sync disabled, two
consecutive operations of the built-in benchmark with quality antialiasing at
Normal and Extreme levels
·
F1 2012 (DirectX 11) – update 9, we
used Adrenaline Racing Benchmark Tool v1.0.0.13 with “Ultra” image
quality settings during two laps on Brazilian “Interlagos” race track with 24
other cars and a drizzling rain, “Bonnet” camera mode
·
Hitman: Absolution (DirectX 11) – version
1.0, built-in test with Ultra image quality setting
If the game
allowed for recording the minimal fps indicator, they would also be added to
the charts. We ran each game test or benchmark twice but only when the
difference between them didn’t get over 1%. If it exceeded 1%, we ran the tests
at least once again to have the results repeated.
