Hoping AMD starts getting all together and get over Bulldozer 'fiasco'.

AMD is behind in 2 ways:

1. Processor Instructions / Specialized sub processors which AMD quote "Does not believe in". So a trans-coding sub-processor like on the i5 doesn't make sense even i5 rip bulldozer apart in trans-coding benchmarks haha. The only thing I will say about this category is if you look very closely at specific benchmarks, Intel has a software/instruction/subprocessor advantage over AMD on top of the hardware advantage which crushes AMD more then it should had they at least caught up on the software / sub processor side of things.

2. Hardware Advantage: Ring-Bus + 3D transistor.
The ring bus is the reason the New Celerons don't suck.
I remember how much Celerons used to suck, and they still do, but they can actually be suitable now days for "normal"
computing.

3. Design: The bulldozer architecture is not very fluid; its a hard multicore design. For example, you can't shift the entire performance of 8 cores unto a single core if the app only uses one core. In this way, the design lacks fluidity / stem cell. As a result, many apps simply won't function to their potential. It would be cool if AMD could hardcode an instruction to force an app to use multiple cores; but much easier said then done.

Summary: Not the best design, but also not squeezing all the juice they could have extracted with more instructions and sub-processors. It is as if, AMD engineers sat in room, came up with proposals, and said, ya this seems like the best one, and went forward without testing whether their theory held in practice, quote "with today's software". Worse yet, no programming language offers any framework to maximize performance on bulldozer. All in all, a major fail. But can never hate competition over Intel's ridiculous prices.