Test Setup and Benchmark Results

Time to put the card to the test. I swapped my good old E8600 for the i950 as a daily rig. For the tests I used to the same ones as in the the Asus GTX 285 Matrix review. The Futuremark synthetic tests opted for are : 3DMark01, 3DMark03 and 3DMark06. For the Game tests I opted to go for games that have an integrated benchmark. This makes my life far easier and I must say that the included benchmark tools work pretty fabulous. I had some issues with Left4Dead, results not really being constant enough .Making it one of the reasons it was swapped for the new Resident Evil 5 fixed benchmark. Sadly my GTX 285 was with another reviewer, so I opted to use the GTX 280 as comparison. MSI claims this card can trounce a GTX 285 when being Overclocked, so it shouldn't have a hard time with the older GTX 280 card.

Here's a rundown on the test rig :

Leeghoofd's i7 i950 Test Setup
CPU	Intel i950 @ 3.8 (200 x 19) 1.15vcore
Cooling	Thermalright eXtreme
Mainboard	Asus Gene II 0904 bios
Memory	6Gb Corsair PC12800 Dominators at 1600mh 8-8-8-24
PSU	OCZ ModStream 700W
Articles linked to this article	Corsair Hydro Series H50 CPU Water-Cooling Kit Review Intel Pentium E6300 Best bang for the buck ? Cache test revisited. DDR3 Roundup: New Elpida Kits from OCZ, Mushkin and Corsair

Windows XP with SP3 was used as OS (will swap to Windows 7 soon). Latest WHQL Forceware drivers from Nvidia's website were used : namely the 191.07 one. Keep in mind that the main objective is to compare the potential of this MSI card here, this is by al means not intended to be a common graphic cards roundup (you've read too man of those already).

Game tests run at 1280 x 1024 and 1920 x 1200 resolution with the integrated benchmark tool, High detail settings selected
Farcry 2
Tom Clancy's HawX
Resident Evil 5 Fixed Benchmark

We took the average FPS from three game runs (all after a fresh reboot). Maximum FPS means nothing if your minimum FPS drops too much, making a game unplayable. Since I'm used to playing First person shooters my own personal minimum limit is 60 FPS. If I can't keep that steady I'm lowering detail levels or if I can sell my wife I buy new hardware.




Small reminder again : the daily clocks are with slight fan speed increase ( 60 vs 40% ). The OC'ed results with max voltage, fan speed at 100%. Without any voltage the card just goes up a bit , but nothing spectacular. With +200mV we get a very healthy increase of 75mhz on GPU clocks and 157Mhz on the shaders. The RAM follows nicely and tops at 1250mhz.



The card scales well in the synthetic benchmarks with the extra Mhz obtained due to the fan speed increase and/or extra GPU voltage. Especially 2003 get's a very beefy boost. The GTX 280 can keep up in 01 though looses badly in 03 and 06 benchmark.
This behaviour also reflects in my game test suite.



Games still benefit by the increased card speeds. But there's almost no gain in fps between the stock MSI Lighting card and the extra 20mhz on the GPU core (pretty logic) Note that the ref clocks is a simulated GTX 275 1792Mb by clocking down the MSI card to Nvidia reference clocks



Almost the same picture as with the lower resolution. A small gain in HawX (directX9) and Resident Evil 5 benchmark. Far Cry 2 gains are less pronounced. The GTX 280 just limps behind its younger brother.