LockerGnome reader Shazad asks, “I have a Radeon 6990 and the frame rates aren’t uniform. I have some spare money and I’m considering replacing it with an nVidia GeForce 590 GTX. Do you think this would be a step up in performance?”

Well Shazad, these two video cards are among the top performers on the market today. Each of these cards have a significant amount of power at their disposal and should be able to eat anything the current game market throws at them with enough room left over to dessert. Here are some of the notable differences between the cards:

Noise
The AMD Radeon 6990 won’t win any awards for its quietness. While it idles at a reasonable quiet volume, the fans pick up once the GPU is under load.

The nVidia GeForce 590 GTX features a single-fan cooling system that blows air over each processor through channels on vapor chambers on either side. Unlike a more traditional solution placing a fan directly over the chip, this one takes advantage of constantly moving air to disperse the heat more efficiently. Previous models from nVidia included heat pipes that worked well to cool at the time, but restricted air flow across the chip requiring a higher fan speed.

Memory
Both of these cards feature GDDR5 memory. While the Radeon 6990 appears to have the edge based on their 4GB of RAM vs. 3GB on the 590 GTX, this doesn’t really tell the entire story.

The Radeon 6990 offers a dual 256 bit memory bus while the nVidia GeForce 590 GTX boasts a larger 364 bit (64×6) pathway. This means the path between the graphics processor and video RAM is wider, allowing more information to flow at once. This can have a direct impact on performance.

The Radeon 6990 has a higher default memory clock at 2500MHz versus the GeForce at 1707MHz. This number can be overclocked through the drivers, though it’s usually easier to overclock the on-board CPU rather than the memory. The memory clock is multiplied by the bus to determine the overall memory bandwidth.

When it comes down to it, the nVidia card has the edge with a slightly larger memory bandwidth of 327.744 GB/sec versus the Radeon’s 320 GB/sec.

GPU
Both of these cards feature dual GPUs on-board. There is a clear difference in terms of standards compliance and processor clock speed.

Both the Radeon and GeForce cards are DirectX 11.0 and Pixel/Vertex Shader 5.0 compliant. The Radeon card, however, is compliant with OpenGL 4.1 where the nVidia card is only compatible with up to 3.2.

The 6990 also has a higher TMU rating at 192 versus the 590 at 128. This rating determines how quickly a card can take a texture and place it around a piece of geometry. For example, the texture of a basketball around a sphere.

At the end of a pixel’s journey, is the raster operator, which adds anti-aliasing and other effects prior to sending the image to the screen. Here, the nVidia 590 sprints with a rating of 96 versus the Radeon’s 64.

Summary
The switch from the Radeon 6990 to the nVidia GeForce 590 GTX may yield better results performance-wise in most areas except texturing. If you are playing a game with a lot of large textures and your quality settings are cranked at full, you may notice a slight difference in performance. On the other hand, the 590 offers a quieter experience with a slight performance edge in some situations.

nVidia GeForce 590 GTX

Source(s):
nVidia
AMD