Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

In a modern fairly powerful system (i.e. i7, lots of power, lots of ram, fast HDD etc) what are the benefits of running two identical graphics cards (i.e. the GTX560) vs running one graphics card?

I know there are benefits for very high-end games, but is that the only place you will see a benefit, or is there any benefit to other grapics processes (editing video for example)?

share|improve this question

The main benefit really is for gaming.

This bring said, there are applications such as video converting/transcoding which are Cuda or OpenCL enabled where the speed can be vastly improved with better/more gpu speed.

share|improve this answer
Gaming, large numbers of monitors, and GPGPU. That's pretty much it. – David Schwartz May 7 '12 at 5:16
More heat. More power! – surfasb May 7 '12 at 5:46
@Surfasb - and much more cost... I remember testing my 8800GTX and worked out that if I left my machine on 24x7, even just at idle, it would cost £65-£70 a year in electricity. – William Hilsum May 7 '12 at 7:05
@WilliamHilsum is correct. Although the benefits of multiple graphics cards do not scale well. Having 2 cards does not mean you will get 2x the graphics power. An average best is usually 50% better performance. – Keltari Feb 13 '13 at 17:17

As far as gaming goes, there would be three reasons...

  1. You wish to run one monitor at the highest resolution and max settings with no lag in performance.

  2. You want to run a ATI Eyefinity or NVidia Surround setup (3x1 Monitors ie 5760x1080).

  3. You want to run a game in 3D mode. Since 3D essentially requires 120Hz monitors and will basically be doubling the requirements of a graphics card since it's showing two images, you'll need more power.

Also, keep in mind that you will want to check the graphic card's potential, motherboard PCIE lane bandwidth. For example, if you have a video card that will push a PCIE2 lane up to 8x, and when using your motherboard in SLI/Crossfire, it will run at x16,x4 then your first card will be used at it's full potential but your second card would be lacking. There's few cards that would push x8 or even x16 so you probably don't have to worry about it too much, just something to consider.

Side note: To run 3 monitors on NVidia, you'll need two cards. Two monitors plugged into the first card and one monitor plugged into the second card. With ATI's Eyefinity, you can run one card and three monitors. Many of the ATI cards will have a combination of 1-2 DVIs and 1-2 Display Ports. The HDMI port on the ATI card is rather useless in Eyefinity setups.

share|improve this answer

In the realm of professional 3d graphic design software, like 3ds max and Maya, renderers started using the processing power of GPUs through Cuda or OpenCL. Using the Gpu simultaneously with the CPU to calculate photorealistic lighting could even triple the speeds of calculation - a medium powered card is as good as an extra i7. Generally the more graphic cards the merrier; Iray, Vray and octane already use GPUs for rendering.

The current problem with that technology is that GPUs have limited memory, an architectural scene could use easily 5 or 6 gb of memory. Only the quadro 6000 has that amount of memory available.

I added two 660 gtx (no sli) under my quadro 4000 and i managed to drop the time required to render one photorealistic image from 5 minutes to 1. It did 4:48 minutes on an overclocked six core 9370k at 4.2ghz ! 1 minute when i added the 3 cuda cards.

This is not applicable in gaming performance or applications that don't us cuda or openCL.

share|improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .