Originally Posted by megahmad
I have read in a website that the sb-e is about 4% faster clock-for-clock than sb, is that true?
Here is clock for clock performance as benched at [H]. As you can see it is slightly slower in all three benchmarks.Source
In general, I would think the above is true (i.e slightly slower than SB) because of the larger cache latency (due to the increase of cache size to 15MB from the 8MB on the 2600K) and the complexly related to two more cores. However, if a certain program/bench loves big caches then you might see SB-E pull ahead clock for clock. Again, this is really minor stuff. I would not think any of it would be noticeable during normal use.
Also, don't forget about clock speed. SB should clock further due to less power draw and way less heat output all else being equal (i.e. same cooling setup). If you remember the original SB reviews, many reviewers had their chips OC'ed to 5Ghz or above, whereas if you read the SB-E reviews most show an OC of 4.6 -4.8GHz. That should tell you somehting about likely 24/7
If you compare a SB chip running 100-200MHz faster than SB-E... then any small clock for clock improvements will be lost and you would only see an advantage in games that are coded for more than 4 cores (very very few games meet this requirement) .
-- Since it does not seem to be relevant to the OP, I am neglecting the PCI-E bandwidth advantage of the SB-E platform which may
lead to benefits in 3/4 GPUs and 3 screen environments (waiting for reviews with this type of setup). For now I will say that in that extreme case you may see gains in a large number of games with SB-E.Edited by JCPUser - 11/20/11 at 6:33pm