Thanks for the education, although I’ll have to research the meaning and function of a few of the above terms.
With almost any aspect of a GPU, it’s pretty much “a number on a box” and, generally, the bigger the number the better the performance.
It’s not just the manufacturers of crappy GPU’s that promote RAM above all other aspects, the big corps (ATI and nVidia) also do it (at least in their packaging—which is pretty much the only thing most consumers will see). Why then, do the manufacturers of quality GPUs choose to promote an inferior, less important aspect of performance? It seems to me they would generate more sales by marketing the more important aspects such as the ones you list above.
To tell you the truth, I don’t know what half that stuff does either
I may be wrong, but I believe the main reasons why GPU RAM is so heavily advertised are that:
A) The general public is already familar with the importance of system RAM on their main computer. Who the heck knows what a Texture unit is, when general GPU architecture completely changes every few years or so?
B)it’s one of the easiest thing to change from the GPU reference design (after a factory overclock). It’s really hard to add additional stream processors or ROPs to a card without a complete redesign, but slap another 512MB of memory on there? Easy. So the higher-end versions of the same cards start getting more memory (with good reason, as there is a performance difference between the two, especially as you go into higher resolutions), and advertising it on the box. The general public latches onto this, and starts associating more RAM with more performance. It doesn’t take long before the GPU marketers catch wind of this, so the manufacturers start releasing $50 cards with “1GB RAM!” and “EXTREME GRAPHICS!” because people who don’t know any better continue to buy them.
