All I hear people talking about now is the new bulldozer and how it will compare with sandy bridge. People keep saying don't build a new PC now just wait for bulldozer. My question is, as a gamer, why should I care? As far as I understand it, all it does is pretty much put 2 cores on what was previously the L2 cache and other dedicated resources and make them shared resources, thus increasing the number of "cores", which aren't "true cores" anymore. Since most games don't even use four cores, why would I even want the increased core count for my gaming machine? Or are their other features that I'm leaving out?
you should wait.Because in the world of technology nothing is predictable.
AMD BD architect is more focussed on "throughput" which is the output of processed data of every core combined but intel is more based towards maximizing processing output of each core.
Wouldn't a regular 4 core processor be better for overall throughput then a bulldozer 4 core since each core has its own dedicated resources compared to bulldozers shared resources between cores?
Data comes in forms of streams through which we call frequency as each frequency hits we get a bunch of data to begin processing.
then we have branch prediction that tries to assemble the data pipeline according to the best and fastest order the processor can begin its work,branch predictors are used before processing any data.However we have a hit and miss rate.
many many more.
This is the only basics that you should know about processing, it takes hours with the help of millions of people to assemble and develop a processor.
I myself think AMD cant beat Intel however as I said anything is definitely possible.
as for ur question "why u should wait"
holding off on ur purchase increases ur options once the product is actually out. If its better u can go for it. If u still decide to go with present products u might at least see price-cuts. over all its always better for the consumer to wait if new product releases are right around the corner.
As far as "should I wait?" goes, you should if you want an AMD system. If it doesn't matter than a core i5 system would give the best performance right now.
It doesn't matter really if bulldozer will beat sandy bridge or not, either system will be quite capable if paired with the right components. A quad core at 3.0 Ghz or higher with a good graphics card will be quite sufficient.
Now if you want bragging rights...We'll just have to wait and see.
I would say waiting only if you plan on going multi-GPU, since since you get more lanes (and USB 3.0 and a few other goodies) with AMD's 900 series chipset, but that will depend on bulldozer pricing.
Although I am personally waiting on bulldozer, I don't think it is the best option for most people. Beyond bulldozer cost being unknown, the 2500k is just plain good enough that it won't likely be your bottleneck: even if BD was 2X as fast as SB, what difference would that actually make to you? Also, Ivy bridge has already been announced to be pin-compatible with SB, so you do have an upgrade path (Recently a big AMD selling point)
Some people may say that the BD won't have much over current AMD chips per core, but comparing an 8core/4mod BD CPU to a 6core Phen2 isn't going to have much difference, but it will be more future proof and BD does support newer instructions like AVX which will kick the crap out of SSE once games/etc support it.
So, just because of AVX, BD will be about twice as fast as current AMD chips.
Anyway, next year, AMD plans on having up to 20core BD chips. I'm sure if you compare a 20core BD against a 6core Phen2, you will notice the BD is a bit faster.
If you're worried about the next 6-12 months, just get any quad+ core that's out right now, but if you're worried about 2-4 years, wait for the BD or Ivy Bridge. A 6 core Sandy Bridge would also last quite a while, but those still carry a premium.
The old "regular" core designs are dead. Neither AMD nor Intel will use them anymore for high end. All your will see is hyper-threading or BD modules as they are a much more efficient use of transistors.