How We Test
Test System Components
All of the tests run on this card were performed using our reference system. Each reviewer for Tom’s Hardware has the same setup, which consists of an Intel Core i7-5830K running at 4.2GHz, paired with 16GB of Crucial Ballistix Sport DDR4 and a pair of 500GB MX200 SSDs. The board we use is an MSI X99S Xpower AC, and the system is powered by a Platinum-rated 850W power supply from be quiet!. All of the components are installed into a Lian-Li PC-T80 test bench.
Software & Drivers
Nvidia is marketing the card as an upgrade for gamers running a GTX 650 or less. We’ve included a GTX 650 in this test to show how much of an improvement the GTX 950 would be. We’ve also included a GTX 750 Ti and a GTX 960. The 960 is obviously a higher-end board, but it shows what some extra money will buy. The 750 Ti is included as a baseline to show what the 950 offers beyond it. All of the Nvidia cards are tested with the same pre-release GeForce 366.65 driver.
The direct competition for the GTX 950, in terms of cost, is AMD’s R7 370, and the next step up is the R9 380. We have samples of both on-hand, so they were added to the comparison. AMD's boards are tested using Catalyst 15.7.1
|Direct X||DirectX 11|
|Graphics Drivers||XFX Radeon R7 370 2GB Black Edition: AMD Catalyst 15.7.1 Sapphire Radeon R9 380 ITX: AMD Catalyst 15.7.1 Gigabyte GeForce GTX 650: Nvidia 355.65 driver Gigabyte GeForce GTX 750 Ti: Nvidia 355.65 driver Asus GeForce GTX 950 Strix: Nvidia 355.65 driver Zotac GeForce GTX 960 AMP!: Nvidia 355.65 driver|
|Middle-earth: Shadow of Mordor||Built-in benchmark, 40-sec Fraps, Very High preset|
|Battlefield 4||Custom THG Benchmark, 100-sec Fraps, Ultra preset|
|Metro Last Light||Built-in benchmark, 145-sec Fraps, Very High preset, 16x AF, Normal motion blur|
|Tomb Raider||Version 1.01.748.0, Custom THG Benchmark, 40-sec Fraps, Ultimate preset|
|Far Cry 4||Version 1.9.0, Custom THG benchmark, 60-sec Fraps, Ultra preset|
|Grand Theft Auto V||Build 350, Online 1.26, In-game benchmark sequence #5, 110-sec Fraps, FXAA: On, MSAA: off, Texture Quality: High, Shader Quality: Very High, Shadow Quality: Normak, Reflection Quality: Normal, Water Quality: Normal, Particles Quality: Very High, Grass Quality: High, Soft Shadows: Softer, Post FX: High, Anisotropic Filtering: 16x Ambient Occlusion: High|
My sound measurements were taken in the same manner as the GeForce GTX 960 reviews that I’ve written this year. The test isn’t as detailed or accurate as Igor's, but I also don't have the same kind of equipment. Instead, I use a hand-held dB meter. It's only capable of detecting sound as quiet as 35 dB. Anything lower than that and the meter reads 0 dB. In the graphs, if there is audible noise that is not registered, the graph will show 34 dB to represent that there is some sound, but an unknown level. Seeing 0 dB on the graph means the card made no discernible sound.
To test power consumption using our reference platform, a bit of creative math is needed. Since Haswell-E processors don’t have integrated GPU cores, we can’t boot the system without a discrete board installed to get a baseline. We are able to estimate consumption based on the approximate power draw of the test bench, though. In our observations, we’ve found that the approximate power draw from everything other than the GPU is 120W. By deducting that from the recorded wattage reported on our in-line power meter, we can calculate the approximate draw of the GPU.
Most of our benchmarks use the same settings as the high-end GPU reviews. However, GTA V and Shadow of Mordor had to be toned down. The rest of the tests are run with the same settings we've used in the past.
All of the games were run at two resolutions: 1920x1080 and 1366x768. If you are buying a graphics card on this kind of budget, you likely aren’t buying a new display. Many older monitors sport that lower resolution, so we thought it appropriate to include it. In fact, Valve’s most resent Steam survey showed that more than 26% of gamers are running at 1366x768.