ASUS GTX 580 Matrix Platinum
Architecture and Specifications
By now we’re very familiar with the GF110 (GTX 580) architecture, so we’ll go with a quick recap. With this generation chip, it’s still based on the Fermi architecture and is very similar to the GF100 (GTX 480) design, though there have been a few changes. The CUDA cores have been bumped up to 512 and Texture Units have also increased slightly from 60 to 64. Polymorph Engines have increased to 16, along with higher core and memory clock speeds. Given these modest increases and nearly identical architecture, it does promise improved power efficiency and temperatures, two things which were sorely lacking in the previous generation.
The diagram below shows the architecture of the GTX 580.
Below is a chart that shows the significant differences in the current GeForce GTX performance lineup as things stand from Nvidia:
As you can see, the GTX 580 is a powerhouse, while the 570 is just a bit lower on the performance ladder. The 560 Ti is even further down, treading more into the performance mainstream market.
Below is a GPU-Z screenshot that shows the particulars of the ASUS GTX 580 Matrix Platinum. As you can see, the clock speeds of 816/1002MHz are very modestt; frankly, we were expecting something more aggressive on such a premium card.
In addition to the lack of a hefty overclock, notice the 1.5GB of GDDR5 memory available; we were a bit surprised not to see 3GB. So don’t expect two of these in SLI to power extreme resolutions in Nvidia Surround; they’ll likely choke with the VRAM lacking. Otherwise, normal resolutions should be fine.
Let’s move onto the package and accessories.