Frame Rating: Eyefinity vs Surround in Single and Multi-GPU Configurations
AMD Eyefinity and NVIDIA Surround plus 4K
As I mentioned on the previous page, Eyefinity and Surround are technologies that have been around for a long time but are picking up steam as the price of high quality 1080p monitors drops. AMD was the progenitor of multi-display gaming with Eyefinity and it has been a part of their promotion and marking for Radeon graphics cards for years, both high-end and mid-range cards.
These configurations put a lot of pressure on GPUs. A 2560x1440 screen results in a pixel count of 3.68 million while a 5760x1080 (triple 1080p screen) setup results in 6.22 million pixels. So, to maintain a steady frame rate of 60 FPS the gamer’s PC must be able to handle 69% more pixel per second. Single GPU configurations (with the exception of maybe NVIDIA’s GeForce GTX TITAN) are unable to really meet this demand while setting games to higher quality settings and thus many users that invest in multiple panels for Eyefinity/Surround are also investing in multiple GPUs for CrossFire or SLI.
And thus our problem is stated. AMD has said that the 13.8 Catalyst beta driver (and the fix found in it) was only good for 2560x1600 resolutions and below. There was some confusion if this meant Eyefinity was included or not as some people speculated that because each screen was only 1920x1080, it would work. As it turns out, that isn’t the case.
Another problem AMD may have coming up is the move to 4K. Current 60 Hz 4K gaming panels like the ASUS PQ321Q use a dual-head configuration. That means that although only a single cable connects to the monitor via DisplayPort, multiple streams are being sent back and forth, reporting to the PC as a multi-monitor connection. Essentially, gaming at 4K 60Hz on current screens is a result of setting up a two screen Eyefinity or Surround configuration with each “head” running at 1920x2160. The combined resolution is 3840x2160 for a total pixel count of 8.29 million pixels, or 2.25x the pixels of 2560x1440.
Obviously multi-GPU setups are going to be demanded for 4K.
The ASUS PQ321Q
In order to test AMD Eyefinity and NVIDIA Surround with our Frame Rating configuration, we had to be a bit more creative. Previous testing was done simply by using a dual-link DVI splitter sitting between the panel and the graphics card, with the secondary output going into our external system and capture card. That card records the raw data and then scripts analyze the video (with overlay) to get performance data.
With Eyefinity and Surround there are three screens so which one do we capture? The answer turns out to be any of them. Thanks to an updated overlay that can present as many sets of overlay bars on the screen as we request, we can capture the left, center or right hand screen to see performance data. We are simply intercepting ONE of the screens rather than the ONLY screen.
Our Datapath DVI-DL Capture Card
As it turns out, performance measurements are the same regardless of which screen you capture from, as you would expect. For my testing then we decided to capture the center monitor which allows to create some side-by-side animation comparison videos as well.
|Test System Setup|
|CPU||Intel Core i7-3960X Sandy Bridge-E|
|Motherboard||ASUS P9X79 Deluxe|
|Memory||Corsair Dominator DDR3-1600 16GB|
|Hard Drive||OCZ Agility 4 256GB SSD|
AMD Radeon HD 7970 3GB CrossFire
NVIDIA GeForce GTX 770 2GB SLI
AMD: 13.8 (beta)
|Power Supply||Corsair AX1200i|
|Operating System||Windows 8 Pro x64|
For my testing I wanted to use the highest end graphics offerings while maintaining price parity. The Radeon HD 7970 GHz Edition has come down in price steeply and currently sells for as low as $370 with a 3GB frame buffer. So while NVIDIA does have faster single GPU offerings, the best comparison for our purposes is the GeForce GTX 770 2GB cards that sell for about $390. I do realize that the GTX 780 and GTX TITAN would offer better single GPU and SLI results for NVIDIA but keeping the prices within range is key for any PC component discussion.