I know but they could easily fit a 3.4 Ghz or even 3.6 Ghz native 6-core in if they raised the TDP envelope to 140w. 3.2 @ 125 is just like the 955 and 965.
Printable View
That's what I would call OC so yes. I think it's a good feature, it boosts where possible. Also, I believe the reason for pro-AMD people to have been talking down the turbo boost on Intel CPUs is because it gives the impression that their CPUs perform much better per clock than an AMD CPU even though the difference isn't that big.
It is likely you can disable the feature -- it won't likely work unless C&Q is enabled anyway. However, it will default to the base clock, 3.6 would likely exceed the thermals. Then, of course, you could just over clock it to 3.6 anyway (black editions are unlocked), which is what most enthusiast will do anyway.
Turbo for an OCer is almost worthless ... most just turn it off and crank all up as high as they can get it.
http://www.tomshardware.com/news/PAX...-X6,10004.htmlQuote:
AMD Brings Phenom II X6, Eyefinity6 to PAX
9:10 PM - March 27, 2010 by Devin Connors - source: Tom's Hardware US
Intel got its six core CPU to market first, but AMD's Phenom II X6 offerings are just around the corner.
While Nvidia might be stealing the show so far at PAX East, AMD is here with some new hardware of their own. While a lot of what was revealed is still under embargo, here's some clarification on what's coming in the next several months.
http://media.bestofmicro.com/,6-F-242583-3.jpg
http://media.bestofmicro.com/,6-D-242581-3.jpg
http://media.bestofmicro.com/,6-E-242582-3.jpg
:)
http://media.bestofmicro.com/,6-F-242583-3.jpg
^
one of the reasons why you wont see much turbo action in most apps... :p:
The task manager shows around 50% CPU utilization or 3 cores,which lines up with AMD's Turbo Core specification that states the Turbo kicks in to 3.6Ghz if <=half the cores are used,meaning 3 cores or 50% of X6 :). You get 3.6Ghz for 3 or less cores used,mind you.And you get a very good 3.2Ghz for 4,5 or 6 cores used which guarantees a big speed up over X4 in multithread and/or multitask scenarios.
Yes if only 3 cores would be used, but windows shuffels the threads between the different cores. so basically it utilizes all 6 cores, and not just 3.
Its the same for intels turbo, thats why you hardly see any situation where you get more then +1x multi.
The only time when i see that happen is, when i assign the task manually to a physical cpu and there isn't any other app running, that utilizes another core more then 1-2%.
I heard that win7 handels this a bit better, but haven't had a chance to test this, since im still on vista.
In the end we'll have to wait and see how it is done, but it doubt it much different from how nehalem handels it aka how windows handels it. Meaning you get a nice standard boost in speed but don't count on seeing the highest multi under normal user behaviour.
According to news
die size of here :
phenom II x6 around ~ 356 mm2 (??)
phenom II x4 around ~ 258 mm2
athlon x4 around ~ 169 mm2
athlon II x2 around 129 mm2
different size of athlon II x2 and x4 is (169-129) = 30 mm2
now it's possibly to make athlon x8 for die 356 mm2 (if we insert two die of athlon x4 in one packet ,, 169 * 2 = 338 mm2 )
http://www.anandtech.com/cpuchipsets...oc.aspx?i=3638
so athlon x6 should be able to handle 90% of performance a Phenom II X6.Quote:
At the same clock speed the Athlon II X4 should offer roughly 90% of the performance of a Phenom II X4.
Hope that Athlon octa core should defeat core i7 !!!
lol, why? 2H 2010 are samples Bulldozer and 1Q2011 maybe a launch. Need not K10.5 8-cores...
according to w0mbat, he could access to the parts that resemble BullDozer
source: http://www.xtremesystems.org/forums/...&postcount=333
yes BullDozer will come out in 2011 , but i don't want to see the dead of AM3.
I read few reports about it being fully compatible.And even JF wrote that bulldozer server parts would drop in to c32/g34 infrastructure.So they have to be electrically comaptible with phenoms.
And then there are multiple AMD roadmaps as this:
http://icrontic.com/uploads/2009/11/...op_roadmap.png
We can be pretty sure that AM3 is bulldozer ready ;-).
X6 listed at giezhals with stepping E0 (Istanbul is D0), no prices yet.
Here's the 960T, showing stepping C2 which must be wrong.
Asus users, here you go
http://i39.tinypic.com/261oq2s.jpg
Gigabyte users:
http://i39.tinypic.com/30lfh2b.png
The problem is that of update frequency and timing, if the windows scheduler switches context over the 6 cores faster than the turbo update can occur, then the turbo algorithm will assume 6 cores are used.
To put it another way, turbo mode does not force cores idle, it watches for idle cores. This is based on the power state of the cores as specified by the OS, if the OS does not let cores or enough cores go into the lower power states, turbo mode does not engage.
Thread scheduling is a component of how well or effective turbo mode will operate as well as the polling/update rate for the method employed. If you recall, Penyrn (mobile) implemented a rudimentary form turbo mode which simply defined new power states for up clocking the cores, the problem was it did not turbo very often and this was because Windows kept throwing out spurious threads that kept the cores awake. There was a big whoopla about this during the Win 7 roll out as I recall, I have done some simple testing myself and frankly Win 7 is the OS to run with this CPU, Vista will perform worse.
Now, having said that, there is no doubt that the X6 is going to be a nice leap over the X4 965 and will be a serious CPU in consideration for anyone building a rig.
i couldnt agree more with that.and considering the price range of these cpu's,they will make for a great centerpiece to build a very powerful system.it will be my next setup,thats for sure,since i cant afford an intel x6 setup:up:Quote:
Now, having said that, there is no doubt that the X6 is going to be a nice leap over the X4 965 and will be a serious CPU in consideration for anyone building a rig.