Friday, December 5th 2008

Phenom II at 6+ GHz: How AMD May Have Done It

A couple of weeks ago, at the AMD Austin Tech Day event, the company stunned the computer enthusiast community, by presenting to the press, the overclocking capabilities of the upcoming Phenom II X4 processor. The enginners managed to boot the machine at CPU speeds in excess of 6 GHz, 6.213 GHz to be precise. A close-up snap of a portion of the CPU-Z window showing the overclock, made its way to AMD's own photostream at Flickr, among other pictures related to the event.

The picture reveals the clock speed at 6213.6 MHz, bus speed at 200 MHz, FSB multiplier at an unreal 31.0x and resulting HyperTransport link speed at 1002.2 MHz. This would mean that they may have dropped down the HT link multiplier (normally 200 x 10.0 for the Phenom II X4 940) to 5x. We already know from previous reports that the vCore was set around 1.90 Volts, and that a copper pot with liquid nitrogen was used to cool the chip. The motherboard used, from the pictures, appears to be Gigabyte GA-MA790GP-DS4H, which is based on the AMD 790GX + SB750 chipset. For more pictures, browse through the photostream. More pictures could be added to the stream once the NDA lifts on December 14.
Add your own comment

88 Comments on Phenom II at 6+ GHz: How AMD May Have Done It

#27
Razr7
then Intel remains The King
It's not about who is the best an any price... 'cause that might make lower performing parts cost more than they should and without competition Intel rips of its customers like it did in the past!
That's meaningful to you of course if you make your own money... but if daddy buys it for you, I can understand how you can like "The King".

Way to go AMD!
10q HTC!
Posted on Reply
#28
cooler
lemonadesoda6Ghz. WOW. Impressive.

BUT, TBH, I really dont know what all the fuss is over Liquid Nitrogen cooled, multipler hacked, one-minute-only computing.

I can get an old Ford pickup, stick an aviation rocket engine in the back, and pull a hell of a stunt... for about 10 seconds before the thing disintegrates.

REAL WORLD computing performance is what wins in my books. Anything else is teenage wet fantasy.

And until AMD can get a CPU on air (or water) in a normal case to beat an i7 on air (or water) in a normal case, then Intel remains The King.
Apparently that chip manage to run crysis couple of time, and it not cherry picked
We've seen Phenom II, Deneb overclocked all the way to 6.2GHz. This was not a screenshot boot, we've seen it with our own eyes as chaps from AMD managed to run the Crysis benchmark several times. The machine was stable on almost every run.


AMD was insisting that this was a stock CPU not a cherry picked one, and that similar scores should be expected from retail products. Intel has never managed to get to over 6GHz and we are sure that Intel is doing some catching up. You naturally need some LN2 to get to such a high scores and it looks that Phenom II might become an overclocker's favorite toy.

This is quite an achievement and with water you will get to 4GHz+, while with a good air cooler you should reach clocks around 4GHz.
www.fudzilla.com/index.php?option=com_content&task=view&id=10801&Itemid=1
Posted on Reply
#29
Drizzt5
If it is NOT cherry picked then that is something truly amazing.
Posted on Reply
#30
HTC
Personally, i would like to see a comparison of i7 and Phenom II @ the exact same speed (around 3 GHz) and with the i7 having both the HT enabled and disabled to see any advantages Intel could take out of it.

After that, OC the CPUs to ... 4.5 GHz or so and repeat the tests so we can see what kind of scalability they offer.

Then, and only then, would we be able to say: CPU "X" is better then CPU "Y".
Posted on Reply
#31
techie81
Let me pull out my Liquid Nitrogen kit that I got at Frys and get this baby going!
Posted on Reply
#32
mdm-adph
HTCPersonally, i would like to see a comparison of i7 and Phenom II @ the exact same speed (around 3 GHz) and with the i7 having both the HT enabled and disabled to see any advantages Intel could take out of it.

After that, OC the CPUs to ... 4.5 GHz or so and repeat the tests so we can see what kind of scalability they offer.

Then, and only then, would we be able to say: CPU "X" is better then CPU "Y".
...and then factor in price. I don't care how much faster an i7 is (and it is, and will be) if I can't afford it.
Posted on Reply
#33
HTC
mdm-adph...and then factor in price. I don't care how much faster an i7 is (and it is, and will be) if I can't afford it.
That's a good point.

Maybe we can ask W1zzard if he can put a "performance per watt and per dollar" in a possible upcoming review ...
Posted on Reply
#34
suraswami
I like that all Red case and 8 panel gaming system. Damn I wish I can spend that much.
Posted on Reply
#35
Binge
Overclocking Surrealism
btarunrYou can do ~4 GHz with air-cooling.
I didn't like the voltages they needed to push 4 GHz though. That really made me :confused: because why would I go after AMD for 45nm process @ 4ghz that's going to obviously burn up. :wtf:
Posted on Reply
#36
mdm-adph
HTCThat's a good point.

Maybe we can ask W1zzard if he can put a "performance per watt and per dollar" in a possible upcoming review ...
It's hard to do -- you have to factor in the prices of RAM & Mainboard to do it.

And then adjust for things like:

Core i7: minus points for prohibitive cost of i7-compatible motherboard, plus points for joing SLI-CrossfireX capability.

Phenom II: minus points for being 2nd place in over speed, plus points for compatibility for current motherboards
Posted on Reply
#37
HTC
mdm-adphIt's hard to do -- you have to factor in the prices of RAM & Mainboard to do it.

And then adjust for things like:

Core i7: minus points for prohibitive cost of i7-compatible motherboard, plus points for joing SLI-CrossfireX capability.

Phenom II: minus points for being 2nd place in over speed, plus points for compatibility for current motherboards
I see your point but what needs to be compared are the CPUs. After we get a "result", then we can factor other hardware costs and see which is more expensive to move to: Intel or AMD.

As for "2nd place in over speed", as you put it, ... says who? Have you seen any benches with i7 VS Phenom II? I know i haven't :(
Posted on Reply
#38
flyin15sec
We can speculate all we want, but until Dec. 14th we won't really know. I'm pretty sure, several review sites already have PhenomII, but can't disclose the benchmarks.
Posted on Reply
#39
Bluesman
BingeI didn't like the voltages they needed to push 4 GHz though. That really made me :confused: because why would I go after AMD for 45nm process @ 4ghz that's going to obviously burn up. :wtf:
At the Austin, TX, test they used between 1.50 and 1.55 cpu voltage. On Crysis with a continuos loop, the temp on AIR was 33C. The peak Gz was 3.9. Temperature is not a problem with Phenom II at 4 Gz.
Posted on Reply
#41
Binge
Overclocking Surrealism
BluesmanAt the Austin, TX, test they used between 1.50 and 1.55 cpu voltage. On Crysis with a continuos loop, the temp on AIR was 33C. The peak Gz was 3.9. Temperature is not a problem with Phenom II at 4 Gz.
Source, cooler, ambient temp, and case? Read up on 45nm process. Voltages above 1.4v is pushing it for that process.
Posted on Reply
#42
HTC
BingeSource, cooler, ambient temp, and case? Read up on 45nm process. Voltages above 1.4v is pushing it for that process.
For Intel CPUs, yes.

Dunno for AMD CPUs because they haven't been launched yet.
Posted on Reply
#43
Chicken Patty
HTCFor Intel CPUs, yes.

Dunno for AMD CPUs because they haven't been launched yet.
I remember when the first Deneb reviews came out, they ran it at 3.5 GHz at like 1.58v or something like that. I think even if 45nm, AMD and voltage get along much better than Intels do.
Posted on Reply
#44
Binge
Overclocking Surrealism
45nm process is 45nm process. I hate feeling like this is an argument because there is proof that no matter what type of electronics you are dealing with the 45nm process will degrade at accelerated (dare I say exponential) rates over 1.4v.

These semi-conductors get smaller and smaller they use less power (and that is good)... there is a point where the voltage can literally change the structure of a gate, breaking it. What I'm trying to say is that a door for people to pass through will hold it's form for just so long with so much use but on Black Friday when the energy of the crowd outside of Walmart got too intense the front doors of the Walmart were forced from their hinges, a greeter was trampled to death, and a pregnant woman severely injured. People still bought goods... Walmart still operated but at what cost?
Posted on Reply
#45
Chicken Patty
Binge45nm process is 45nm process. I hate feeling like this is an argument because there is proof that no matter what type of electronics you are dealing with the 45nm process will degrade at accelerated (dare I say exponential) rates over 1.4v.

These semi-conductors get smaller and smaller they use less power (and that is good)... there is a point where the voltage can literally change the structure of a gate, breaking it. What I'm trying to say is that a door for people to pass through will hold it's form for just so long with so much use but on Black Friday when the energy of the crowd outside of Walmart got too intense the front doors of the Walmart were forced from their hinges, a greeter was trampled to death, and a pregnant woman severely injured. People still bought goods... Walmart still operated but at what cost?
thats a unique, but great example binge :toast:.
Posted on Reply
#46
Bluesman
BingeSource, cooler, ambient temp, and case? Read up on 45nm process. Voltages above 1.4v is pushing it for that process.
Many sources is my reply but here is one SOURCE with a quote:
While I can't give exact numbers in terms of GHz, I can give you some of the voltages and temps seen while running the Crysis demo. Keep in mind these temps are not anywhere near 100% CPU usage, but they aren't exactly idle either.

* Air: Cooler Master Hyper Z600 - Around 1.5-1.55V / 33C
* Water/Tec: CoolIT Freezone Elite - Around 1.6V / 38C
* Phase Change: VapoChill - 1.7V / Below 0C (software / MB wouldn't report negative numbers)
* LN2: 1.75V (at least during the Crysis demo... *grin*)
As you can see in the quote, the AIR run used a Coolermaster Hyper Z600. One XS poster got 4.1 Ghz by changing settings in addition to the multilier. This same poster has also learned from an AMD rep at the demo that the max multiplier is 48X. SOURCE
Posted on Reply
#47
Binge
Overclocking Surrealism
He even says not full 100% usage and that environment is a test bench with most likely excellent ambient temps and no case with plenty of air... :shadedshu
Posted on Reply
#48
Bluesman
Binge45nm process is 45nm process. I hate feeling like this is an argument because there is proof that no matter what type of electronics you are dealing with the 45nm process will degrade at accelerated (dare I say exponential) rates over 1.4v.

These semi-conductors get smaller and smaller they use less power (and that is good)... there is a point where the voltage can literally change the structure of a gate, breaking it. What I'm trying to say is that a door for people to pass through will hold it's form for just so long with so much use but on Black Friday when the energy of the crowd outside of Walmart got too intense the front doors of the Walmart were forced from their hinges, a greeter was trampled to death, and a pregnant woman severely injured. People still bought goods... Walmart still operated but at what cost?
Great analogy Binge! I think the reason you don't get high heat initially above 1.4 with Phenom II is because of very low channel leakage.
SOURCE
AMD's transistors exhibit very low channel leakage. Our transistor benchmarks indicates that leakage current is less than one-third of the value measured on AMD's 65-nm process. It's also significantly lower than the Intel 45-nm HKMG process. In fact the Ion/Ioff ratio for AMD's PFET is nearly 10 times better than that for the Intel PFET.
Posted on Reply
#49
Binge
Overclocking Surrealism
Detailed article, I'm going over the numbers of it myself right now :) Thanks Blues!
Posted on Reply
#50
HTC
BluesmanGreat analogy Binge! I think the reason you don't get high heat initially above 1.4 with Phenom II is because of very low channel leakage.
SOURCE
Interesting read: thanks, dude!
Posted on Reply
Add your own comment
May 21st, 2024 14:29 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts