ајде да спориме тогаш za битовете си стојам на зборот дека е така но не можам да кажам дека е секогаш така земи пример за GF6200 256mb 32bit 460/333mhz во спредба со GF6200 128mb 128bit 380/310mhz (ако си имал искуство со такви работи)
а и нешто од твоја страна се надевам дека знаш англиски:
Most video cards are specified in part by a bit width. It is important not to be mislead by claims from video card manufacturers about how many bits their card "is". Many companies are now coming out with cards that are claimed to be "128 bits" and the implication (or worse, direct statement) is that these are much better than 64 bit cards. This is very misleading for several reasons.
The bit size of a video card usually refers only to one portion of its operation, such as how many bits it uses to access the video memory. Virtually every card, for example, still accesses the PCI or VLB system bus using 32 bits, no matter what the width of the internal processing circuits is on the card. This means a 128 bit card cannot (all else being equal) have double the performance of a 64 bit card. Perhaps more importantly, there are no standard definitions of what exactly a "128 bit card" means. Two different manufacturers can have totally different standards for deciding what is a 64 bit card and what is 128 bits.
A card with a wider internal memory bus can in theory have more performance, but remember that it is just one component in a complex performance equation. Don't be dazzled by the big numbers. A high-quality card that calls itself "64 bits" can be much faster than one calling itself "128 bits" if the former has a better-implemented chipset--or better drivers.
За тоа дека Nvidia не се скапи или што и да си сакал да кажеш со тоа.... сепак сакав да кажам дека пример:
GeCudu или Palit со таков чипсет на графичка можат да се наречат срање графички за overclock за разлика од ATI или Nvidia кои можат со overclock да искараат 30% подобри перформанси