QX6800 vs. X6800

[cb62fcni]

So anand's has posted a decent little overview/review of the QX6800, and I found it surprising that as opposed to the E6700 v QX6700, the QX6800 consistently outperforms the X6800 in gaming applications. This seems like it's more than a simple clock increase, architecture optimization? Anyone got further info?

http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=2963

 

[ajfink]

Nope - just the extra cores coming in handy.

 

[HotFoot]

All I've seen online is that this is a new stepping, and that efficiency has been increased. If there were other optimisations, I don't know.

Really, I can think that the extra two cores ought to handle background tasks better, and therefore give a slight improvement to games that aren't quad-threaded. However, the penalty for the MCM approach may explain why quad-core chips have been slower than dual-core at the same game (except those, of course, that use more than two threads). Still, you'd think the penalty/benefits would be fairly equal across the board, or at least the greatest benefit/penalty ratio would go to the slowest-clocked quads.

At this point the only reasons to choose the dual-core over the quad are if you're on a tight budget (but the x6800 isn't cheap either), want more overclocking headroom, or are up against a wall in terms of how much heat you have to deal with.

 

[cb62fcni]

Yea, I was seriously considering getting a Q6600 after price drops take effect. However, it looks like it won't be able to compete against the QX6800 clock-to-clock. Not a huge loss though, the difference is pretty slight.

I for one am glad they addressed this problem though, I'm sick of people going on and on about how the quad-core gets beat out by dual-core in apps. This news should discourage the quad-nay-sayers.

 

[cb62fcni]

Nope - just the extra cores coming in handy.

That was the point, more cores with the 6600 and 6700 actually decreased performance in some apps such as gaming. The 6800's extra cores don't seem to have the same effect.