well Intel HAS always dominated the cpu market. Intel was the company that invented the cpu so AMD loses right there. ATI beat Nvidia because ATI always had more stream processors per card then Nvidia but their architecture and drivers weren't up to par. It was only a matter of time as I had predicted and sure enough ATI eclipsed Nvidia. AMD is already behind Intel in every respect. Intel has better architecture, lgas, dies, wafers, and cpu features. AMD does not have one area at which they are better like ATI had vs Nvidia. The i7's have been out for about 2 years and AMD STILL can't beat them.
1)Wrong, the entire Pentium 4 era was a complete flop for Intel. The netburst architecture was extremely ineffecient, clock for clock, even an athlon XP was nearly twice as fast per Ghz, let alone the Athlon 64. Just because intel has had a hold in the market for a few years now, that doesn't change the fact of how long they stuck with an architecture that simply "sucked".
2) The nomenclature CPU has been around long before intel ever existed as a company. Intel build the first "CPU on a chip" so to speak with the 4004 however.
3) Wrong, drivers are not any reason why ATi cards have been slower.There is quite a difference between a stream processor used in an ATI GPU and a shader used within an nVidia GPU. Shaders are more powerful and much larger than a stream processor like ATi uses. There is no way to directly compare the numbers, this is why with the GF100 GPU's from nvidia with 512 shaders on die(with some disabled) is still a much much larger core than the HD58xx gpu line.
4)Processors take time to develop, just because the i7 has been out for awhile, AMD had to develop an interim product to bulldozer(the phenom II's). As far as features/instruction sets, bulldozer is going to have the same instruction sets as sandy bridge(SSE 4.1/4.2, AES, CLMUL, AVX) and will have instruction sets that sandy bridge will not have (XOP, CVT16 and FMA4). As far as wafers and processes, current wafer technology between AMD and Intel is roughly identical, bulldozer will be using the same 32nm SOI process that intel will be using for sandy bridge.
Well, the difference with this one is that AMD took two dies that filled the entire square, whereas the C2D and Pentium 4 must have only filled half of it, right? Maybe Cedar Mill filled a quarter, but of course a Pentium Q would be waaaaaay too hot.
No, the cpu die is no where close to filling up the entire chip, in fact, you could probably fit all the magny cours dies onto a single AM3 sized chip, but unlike the previous non monolithic dies used, in essence, each cpu die has a dedicated set of pins for power/data, which is why the socket is nearly twice the size of a standard socket. The octa core processors use two quad cores, and the dodeca core processors have two hex core dies.
Does the manufacturing size of the CPU determine how large the die is, though? For example, could you fit more 22nm cores on a die than 32nm ones?
