Short bio: Computer Scientist, FOSS supporter (read more)
Tux Machines (TM)-specific
Whenever ATI or Nvidia bring out a new architecture, a slew of derivatives invariably follows. The sequence goes like this: First comes the rock-star high-end part with the most vertex and pixel shader units, the highest clock speeds, and all the bells and whistles. These appear in cards priced between $500 and $600, and give cachet to all the parts in that architecture family. Usually, the midrange part follows, and has a reduced number of vertex/pixel shader units, and a lower price-tag, normally from $200 to $250. This is the real workhorse for GPU makers, and makes them the bulk of their money.
There has historically been a desert between about $200 and $500 price-points, but the most recent generation of GPUs—Nvidia's NV4x and ATI's R4xx—have seen the rise of $300 and $400 upper-midrange offerings. These additional parts, plus the AGP-to-PCI Express transition have caused the number of available GPUs to mushroom, giving us choices aplenty. Which GPUs really give you the most performance for the fewest dollars? That's today's topic, so strap yourself in for today's ride through the GPU landscape where we'll show you best (and worst) values for your money.
Full Story. Interesting, great info!