Apokalipse,
LOL !!
I edited my previous post 2 times. In the first edit I added the word always before I even read your reply.....
E6550 has 667Mhz slower clock speed than 6000+, yet it performs as good as 6000+ if not better
which is still less than a 4/3 performance advantage (dividing the IPC of the Core 2 by the IPC of the K8)
http://www.anandtech.com/printarticle.aspx?i=3038
Don't tell me that K8 isn't far behind Core 2
Because you won't listen?
Alright.
Thats your opinion.... but I don't agree with you
Here we go...
First: Die shrink is big improvement. Die shrink allows processors to produce less heat, which means processors can run easily at higher clock and have better overclocking
Die shrink is not an architectural change. It's making what you already have smaller.
45nm quad cores can have Thermal Design Power as low as 50W !!!
as low as. Meaning, on the low-end chips.
Also, 45nm quad core will run beyond 3GHz. The first 45nm Yorkfield that will come out on Q4 2007 will have clock speed of 3.33GHz
And that will have good competition from a 3GHZ Agena, which not only has a 4 IPC pipeline, but also only 12 stages, a better caching system, a faster HTT bus, better virtualisation, branch prediction, prefetching and latencies (especially between the cores)
Second: You claimed that the extra cache is small improvement !! But extra L3 cache that K10 have is small improvement in my opinion.
the significance of the L3 cache is not the amount. It's the fact that it's a third layer, shared between all four cores (on the one die)
When Intel added L3 cache for some of the P4 extreme it barely made any difference..........
The Pentium 4's caching system is not exactly like AMD's is.
It's not just
that it will have L3 cache that will make a difference, but how it is used.
Now, I wonder where exactly your argument is going?
All I said was that AMD will have four cores on one die, and now you're touting the Core 2 as the best thing since sliced bread.