Very interesting, especially the results for SPECint_rate_base2006.
Printable View
I dont care what an Intel favored test shows. No virtualization at all... Wheres the benches showing the FP power of the chip?
It is an early release, lets see the chip in Q4 on a newer platform and we dont even know how it reacts to overclocking yet.
Because they are exercising server workloads, and AMD's dual socket interconnect fairs better than Intel's dual FSB.... on CPU bound workloads, this is not the case... i.e. Desktop.
AMD will be competitive winning some losing some on server applications, and will fall short on DT....
Core for core, clock for clock -- Barcey improved about 15% over K8... this is short of the Core advantage by 5-10%....
Jack
Are you high? They already said it's not going to overtake Intel in a performance war. And my response was to his remark about how it's going to respond to overclocking, meaning that the 30k 06 benchmark from INQ was BS.....Wake up buddy, do you hear that knocking? It's called the hard truth. AMD is not going to get the performance crown back with K-10. Sad really as I was hoping it would, as that would spark a huge price war and we all would be winners.
stop right there. they hype is over. if there is nothing dreadfully wrong/crippled like disabled cores/cache NOTHING MUCH will change. just look at the super duper miracle x38 chipset from Intel. 1-5% improvement, but it uses a lot more power and is really expensive. It's just a chipset after all! :) (not to upset any intel funs: but it will be a hell of a great platform for Intel's penryn, 4ghz penryn with 2ghz ddr3 anyone? it is already set in stone that this setup will own phenom, if anandtech didn't lie to us...)
We can't because it simply ain't true. k10 is a very specialised/unbalanced arch and thus it's slower/much faster then core 2 in different applications. on the serverside it is much better than xeon at its CURRENT price point... but on the desktop it is MUCH WORSE than the currently available Intel offerings. However phenom will improve on that so it won't be *that bad* with a 2,6ghz phenom anymore.Quote:
Originally Posted by BeardyMan
Anyway I agree: overall the K10 is a superior arch than c2d, at least for server workloads. if it only scaled as well and was able to compete with penryn in future. well let's see how things turn out...
:shakes: I guess I'm going to have to go Intel with my next build :( / :)
Ahahahahaha! This is absolutely hilarious that you can believe such things. You really need to open your eyes and clear your head of all the propaganda that's been thrown onto the web with this product launch. K10 looks to be another R600, took forever to develop, didn't live up to the hype, and disappointed the entire enthusiast market. :shakes:
edit: as much as you'd like to say it, the enthusiast market drives the server market.. I'm an overclocker and I'm responsible for major upgrades within my department and I'm not the only one..