2% faster clock for clock SINGLE CORE performance increase? whys it such a large increase with all 4 cores loaded? because if you've run a quad, and run say 1 instance of super-pi, great, 2, not at all bad, basically same performance, 3 instances, very slight decrease, 4 instances massive hit to performance.
THe only increase in performance is scaling on the 4th core/instance of a program essentially. So the increase in speed will pretty much only be visible when all 4 cores are FULLY loaded. Which is ultra rare in anything non synthetic/home use. Encoding is always always going to gain the most by getting more cores. However, games, if they get the same 2% boost due to not maxing out 4 cores, would mean a several hundred pound chip, and expensive board, memory set up for triple channel(ie 6gigs of ddr3) means an incredible cost for very very little difference, if any noticeable difference.
Remember 99% of games right now play the same on a Q6600 at 2.6Ghz as at 4Ghz.
Its not a great chip, architechturally its just different really not faster, its simply a much better scaled setup in terms of core communication and bandwidth, which is great.
AS per usual, people doing heavy rendering work will love it, or encode all the time, but the majority of home users, gamings the biggest strain and I can't see there being a single benifit for a couple years. Cheap quad is where its at for 99% of the population right now, cheap these won't be for a whole setup.
Haven't Dell or HP shown largely due to this scaling, that recent Phenom systems of theirs are outdoing the Xeon quad's. Phenom's aren't bad at all, when they get the speeds up in terms of benchmarking they'll do pretty damn well as they do scale with speed better than your average chip. THe only issue is AMD simply not being able to switch to lower process's as quickly as Intel. 58 fabs, shut one down at a time for Intel to switch to 45nm, tiny production slow down. shut one of 3 fabs down for AMD, at a time they simply can't afford not to supply chips to customers, and they lose 33% of production. Thats the reason AMD aren't on 45nm yet, nothing more, nothing less.
Its the reason they are considering selling all the fabs, so they can simply change production to a different 45nm plant while others upgrade without cost to them. the other choice is building the billion dollar plant in New York they have approved(pretty much). Either way works, but selling the fabs gives them a huge increase in funds short term, while building another fab puts them further in debt, but increases value of company and long term profit outlook.
nice to know lets just hope amd can catch up to keep them prices down
