AMD's 65nm Preview Part 2 - The Plot Thickens (Updated with Information from AMD)
by Anand Lal Shimpi on December 21, 2006 12:12 AM EST- Posted in
- CPUs
Final Words
Now that the pieces are falling into place we are able to understand a bit more about the implications of AMD's move to 65nm. It's clear that these first 65nm chips, while lower power than their 90nm counterparts, aren't very good even by AMD's standards. Already weighing in at the high end of the voltage spectrum, we hope to see more overclockable, lower power offerings once AMD's 65nm ramp really starts up. It's a constantly evolving process and if this is the worst we will see, it's not terrible; AMD can only go up from here, but it does mean that you shouldn't hold your breath waiting for the right 65nm AMD to come along.
Performance and efficiency are still both Intel's fortes thanks to its Core 2 lineup, and honestly the only reason to consider Brisbane is if you currently have a Socket-AM2 motherboard. It is worth mentioning that AMD still has the lowest overall power use with its Athlon 64 X2 EE SFF processor, but in terms of performance per watt efficiency it's not all that great. We would really like to see an EE SFF successor built on AMD's 65nm process, but we have a feeling it will be a little while before we are graced with such a delicate creature.
The step back in performance with Brisbane is truly puzzling; while none of our individual application benchmarks showed a tremendous loss in performance, it's a very unusual move for AMD. The last thing AMD needs to do is take away performance, and based on its current roadmaps the higher latency L2 cache makes no sense at all. Either AMD has some larger L2 cache variants in the works that we're not aware of, or AMD's cache didn't take very kindly to the 65nm shrink. As soon as we get the official word as to why L2 access latencies jumped 66% with Brisbane we'll be sure to report it; until then we can only wonder.
We long for the good old days, when a die shrink meant ridiculously overclockable processors, back before a die shrink was coupled with a sneaky decrease in performance. While Brisbane is far from a Prescott, it's not exactly what we were hoping for from AMD's first 65nm Athlon 64 X2. Hopefully they can work out some of the process' kinks in time for the K8L launch.
52 Comments
View All Comments
theteamaqua - Thursday, December 21, 2006 - link
man i hope this thing overclocks like conroe.... otherwise no one will get quad fatherbut i already have E6400 @ 3.4GHz ...
might get Q6600 , Q6400 or Yorksfield or Altair ... ill what see what happens
clairvoyant129 - Thursday, December 21, 2006 - link
Right here,http://www.anandtech.com/cpuchipsets/showdoc.aspx?...">http://www.anandtech.com/cpuchipsets/showdoc.aspx?...
Nothing special though.
Better to stick with the 90nm X2 then this piece of junk.