Sign in with
Sign up | Sign in

Test Settings

MSI Eclipse Plus: Does nForce 200 Boost 3-Way SLI?
Test System Configuration

Intel Core i7-920 (2.66 GHz, 8.0 MB Cache)
Overclocked to 3.60 GHz (18x 200 MHz), 1.35V Core

CPU Cooler

Swiftech Apogee GTZ Liquid Cooling

(PCIe 2.0, x16/x16/x16)

MSI Eclipse Plus, BIOS 1.0 (03/31/2009)
Intel X58/ICH10R Chipset, NF200 PCIe Hub, LGA-1366

(PCIe 2.0, x16/x8/x8)

Asus Rampage II Extreme, BIOS 1306 (04/09/2009)
Intel X58/ICH10R Chipset, LGA-1366

(PCIe 2.0, x16/x16/x4)

Asus P6T, BIOS 0306 (02/03/2009)
Intel X58/ICH10R Chipset, LGA-1366


6.0 GB DDR3-1600 Triple-Channel Kit
CAS 8-8-8-24 at 1.65 volts


3 x EVGA GeForce GTX 285 1 GB SLI
648 MHz/1,476 MHz GPU/Shader, GDDR3-2484

Hard Drive

WD VelociRaptor WD30000HLFS
0.3 TB, 10,000 RPM, 16 MB Cache


Cooler Master RS850-EMBA
850W, ATX12V v2.2, EPS12V


Lite-On LH-20A1L, 20X DVD±R


Integrated HD Audio


Integrated Gigabit Networking


Operating System

Microsoft Windows Vista Ultimate x64 SP1


Nvidia GeForce 182.08


Intel INF

Several games are at least somewhat performance-limited by even the highest-speed CPU. Overclocking a Core i7 920 to 3.60 GHz reduces its impact on this graphics benchmarks. That’s equivalent to two model-speeds above Intel’s fastest-available LGA-1366 processor.

Swiftech’s Apogee GTZ water block works with the company’s triple 120mm fan radiator and MCP-650B water pump to keep CPU temperatures within 10° Celsius of ambient at 100% processor load.

The point of today’s article is to test for potential bandwidth limits on the second and third slots of standard X58 motherboards, whether in x8-x8 or x16-x4 mode, compared to the nForce 200-enhanced MSI Eclipse Plus. Borrowing three EVGA GeForce GTX 285 graphics cards from another test system made this possible.

Benchmark Configuration

3D Games


Patch 1.2.1, DirectX 10, 64-bit executable, benchmark tool
Test Set 1: Very High Quality, No AA
Test Set 2: Very High Quality, 8x AA

Far Cry 2

Far Cry 2 Benchmark Tool
Test Set 1: Ultra High Quality (DX 10), No AA
Test Set 2: Ultra High Quality (DX 10), 8x AA

Tom Clancy's H.A.W.X.

Game Version 1.00, in-game performance test
Test Set 1: Highest Details, DX10, SSAO, No AA
Test Set 2: Highest Details, DX10, SSAO, 4x AA

Left 4 Dead

Game Version, custom timedemo
Test Set 1: Highest settings, No AA, No AF
Test Set 2: Highest settings, 4x AA, 16x AF

Stalker: Clear Sky

Clear Sky Benchmark File Version
Test Set 1: Ultra preset, DX10, No AA
Test Set 2: Ultra preset, DX10, 4x AA

World in Conflict

Patch 1009, DirectX 10, in-game performance test
Test 1: Very High Details, No AA / No AF
Test 2: Very High Details 4x AA / 16x AF

Synthetic Benchmarks and Settings

3DMark Vantage

Version: 1.02, GPU and CPU scores

Having this much graphics power to draw from eliminated any need to benchmark at medium or lower details, so we maxed everything except anti-aliasing and anisotropic-filtering for the “low settings” test to gauge playability, then cranked up AA and/or AF to further stress slot bandwidth. Tom Clancy’s H.A.W.X unfortunately won’t run at 2560x1600 with 8x AA, forcing us to choose 4x instead.

React To This Article