Kentsfield & BadAxe2 > @3.7GHz On Reference Cooler!

I'm looking forward to the QE6600 in January 2007. It should overclock nearly as well but be a good deal cheaper.

Bad Axe 2 is supposed to be launched when Kentsfield launches.

Improvements over Bad Axe:

1) Official DDR2-800 support
2) Should work with DDR2-1067
3) Improved margin on the FSB so better overclocking
4) New SATA RAID controller from Marvell
5) Improved Overclocking controls in BIOS

Legit reviews from IDF has a mini review of the Bad Axe 2 Fab D. Production board will be a Fab E.

http://www.legitreviews.com/article/395/1/

1. Kentsfield/Conroe/XE Support ECO’d in BX WW13 (AA# -304)
- New Heceta to support PECI
- Layout and stack-up changes/tuning that gain margin for both default and over-clocked operation
2. SATA II Controller
- Marvell 4-port SATA2 PCIe controller
3. Sigmatel Audio Upgrade (for Vista compliance)
4. CK-505 Clock
5. Physical switch for Power On/Off
6. Enhanced voltage tweaking options for non CPU related circuits
7. BIOS Enhancements
- Improved User interface/OC steppings

I love my AMDS, but the fanboys truly are pathetic.

Lets do some simple math (apparently beyond the capacity of the "horde" )

1 x 4 = 4
2 x 2 = 4

Hmmm, 4 = 4, no "glue" required

4x4 is not designed to compete with Kentsfield. Neither is Barcelona, but Barclona's specs indicate a theoretical 200% increase in FP and 100% in integer. If their new prefetch and 256bit L1 comes through integer maybe increased 150%. If the "improved IPC" core touted in the Barcelona pics



gets to a sustained 4 IPC, that translates to 25%. Don't count AMD out because the largest CPU manuf finally has a modern chip.

But in your last 1500 posts you are trolling about how 4x4 will outperform Kentsfield and cost less. Are you admiting your mega stupidity?

Idiot, K8 can perform 1 128bit FPi per clock, K8L will perform 2 128 FPis per clock, while the ALU of K8L will remain to be the same as on K8. That is 100% increase in FP and 0% in integer, you stupid troll.
The 256bit L1 and the new prefetcher are required becouse of the 100% FP incerease, stupid troll.

You don't know idiot, what are you trlling about idiot.

Moron, depends what is "modern chip" for you, bnut if you are talking about the performance AMD needed 20 years to produce x86 chip that can stand shoulder to shoulder with the x86 chip of the largest CPU manufacturer.

Jack,

Dont take this the wrong way, but I dont think your going to find a "Kick the Kick Baron" habit support group here at THG.

You're on your own in that endevour.

I can't wait.

Jack,

Dont take this the wrong way, but I dont think your going to find a "Kick the Kick Baron" habit support group here at THG.

You're on your own in that endevour. It's like when you get a new puppy, and he pees or poops on your rug. You have to rub his nose in it, give him a smack, and say......bad dog!!!. Then hopefully some day he'll figure it out. Baron needs to be trained to leave his sh*t outside.



Ahhhhh....If only that technique would work with Baron......[/wistfull contemplation]. Most pups my family has had have learned that lesson in significantly less than Barons 9+ months here at THG forums. I wonder when we can move on to the more advanced commands, like 'sit' and 'heel'

4x4 is not designed to compete with Kentsfield. Neither is Barcelona, but Barclona's specs indicate a theoretical 200% increase in FP and 100% in integer. If their new prefetch and 256bit L1 comes through integer maybe increased 150%. If the "improved IPC" core touted in the Barcelona pics



gets to a sustained 4 IPC, that translates to 25%. Don't count AMD out because the largest CPU manuf finally has a modern chip.

How many times must I tell you fanboys. K8L is not Barcelona. The picture you just posted is of Deerhound/Greyhound (Server/Desktop K8L Variants).

Barcelona is nothing more then four K8 cores with a die shrink to 65nm.

K8L variants (Deerhound/Greyhound) won't appear until the 1H08 (1st Half of 2008).

Til you're blue in the face....at the very least.


Which, by no smal coincidence, is at least the same amount of time the average consumer is going to have to wait since the "official" release of X2 5200 and the time the can actually get their hands on one.

Makes you wonder with all the promotion AMD's been doing for this wonderous new "intel slaying" technology when the average consumer will actually be able to get some.

Yes and theoretical numbers are always prettier then real world numbers. Why even bother bringing up theoretical untested performance ?. Besides that fact amds a$$ is going to be a sore for a week after that beating the kentsfield just lashed out.