The new versions of the GeForce2 have not been changed by no means into his visual capacities (T&L, TwinView, etc.); it is more, they are chips completely compatible with the designs of cards GeForce2 previous MX.

The only introduced differences have been in the speed paragraph of the chip and graphic memory. The following table compares some of the characteristics of the chips the most common NVIDIA destined for the current market of average scale and average - fall:

We might give more information, differ according to the number of applied textures, etc., but with this we can do an idea to ourselves where from the differences take root. Since we see, at chip level three GeForce2 MX are almost identical: the MX and the MX 200 go at the same speed, and the MX 400 slightly more rapid... which practically does not matter, how we will see later.

What he emphasizes is the memory paragraph: the MX 200 "cuts away" the breadth of bus of graphic memory to only 64 bits, therefore the final memory band remains in a few almost ridiculous 1,33 GB/s (64 bits x 166 MHz / 8 bits for every byte).

Also, this value we cannot multiply it by two by means of the technology DDR, since it is supposed that only it uses memory SDRAM. For his part, the MX 400 is only a "classic" MX with overclocking official. "Overclocking" to the effect that his chip and the memory that it uses go more rapid, but of course with all sorts of guarantees, eye.

Bearing in mind the behavior observed in the modern games by most of the graphic chips of the market, including those of the family GeForce, this enormous change in the memory band implies that the MX 200 is MUCH WORSE than the MX or the MX 400. We are going to see it...

As we were saying, on the market some cards meet already these chips. Nevertheless, it seems that for the time being the most common models include the MX 200, perhaps because we have not seen scarcely mark solutions, but rather "generic" cards.

Therefore, we have chosen one of these cheap cards and "without well-known mark" like exponent of the MX 200. A very important thing to be born in mind is that in this class of simple graphic cards, with poquísimos components, with the chip made by the same NVIDIA (good, really for the Taiwanese TSMC) and based in most cases on the excellent reference designs and drivers of NVIDIA, the yield of a generic card is identical to of one "of mark". Anyway, that it is difficult them to harm.

Another thing would be if that of mark was including more rapid memory (the memoirs of the previous table are the recommended ones, which in practice can change in a few tens MHz), but this will not be hardly ever the case with a MX 200.