Phenom II X4 940 > 4ghz++

gr8... im waiting for deneb processors..looks really hot & a kickintheass for INTEL ... WOHOOOOO GOOOOO AMD !!!

We both are I'm confused as to what this is attempting to amount to.

Word, Playa.

Does that overclock comparison come from AMD? LMAO, I'll love it if its tru but until I see reviews...

What is the asterisk after Overclock mean?
I'd like to see what the explanation of the asterisk is for. Is it warning about overclocking, or is it to explain to use a specific type of HSF for that specific overclock?
They explained the other asterisk about the 800 series using only 6MB of L2 and L3 cache. So, why no explanation for the overclock asterisk?
Also, is that the TDP rating being used? I thought AMD used the ACP rating.

If these are even in the ballpark i'll think of getting one. I could do with a CPU upgrade, but the present crop of AM2+ chips just aren't worth dishing out any cash on.

I'll just say, look a little closer at the slide.

[edit]

At the 2 bullet points above the "MASSIVE HEADROOM" for overclocking bullet point.

pie was faster


this time...

I see.
So, they are basing this overclock result on those 2 applications being used, as well.
But couldn't the 9950 also use the same ACC and Overdrive applications? There has been 9950's overclocking higher with the newer SB750 chipset. So, why not use the ACC and Overdrive with the 9950 as well? Or does the ACC and Overdrive only work with the Phenom II?

I'm not familiar with AMD setups, so I am just trying to understand the marketing being used here.

I don't think that the Deneb cores will be anything faster than what has been achieved with the latest chips from Intel, but if I were going for a new pc setup sometime later next year(or something like that), I would be far more interested in the deneb cores than what has been released from Intel.

I earlier heard about stability problems when the Phenoms was released, especially when old Windows games was executed and in this regard I'm concerned about how good the Phenoms really are.

Anyone here who has tested the Phenoms and know if they worked good with older Windows games?

I have heard good things about Phenoms too and I suspect that Linux would, perhaps, treat it a bit better than what Windows has had and still does. And when I'm saying Windows here, I'm talking about Windows XP.

well yeah, acc and overdrive work on the current phenoms as well... but I don't think this whole thing is necesseraly that complicated. They are just trying to market their software along the hardware and/or trying to show how much better the new phenom II will be than the current ones.

and if you look at the slide from top to bottom, the word overclock on the diagram comes 'first' on the right hand side. or something...

ACC is not an application. It is a bios setting.

Overdrive is a software application that allows you to change things without rebooting. You can still overclock using the bios. (Which is nearly always better anyway.)

Okay. So, ACC is a BIOS setting and Overdrive is software based.

But that doesn't explain how the Phenom II 940 has a seemingly endless "headroom" to overclock, based on the chart shown, when using the same technology that is available for the Phenom, now.

Is all this headroom available simply due to a shrink to 45nm? It's definitely not due to power usage, especially considering that the TDP of the 940 is 125W, and the 9950 had a TDP that ranged from 125 to 140W. So, 0-15W difference and 45nm shrink equals unlimited overclocking headroom? That's a bold statement to make, and one that will be challenged, I predict.

Definitely a bold statement. I don't believe it for a minute either. I am guessing it to be on par with the Kentsfield, maybe a slight edge. They will gain some headroom for overclocking on the die shrink but not some kind of magical, no ceiling, no thermal issues, fairy tale.

Yes, the unlimited might be a stretch...but when did you ever see a CPU company tout it's CPU's overclocking ability? Or add a chart that ends at 4GHz, yet the overclocking arrow goes on beyond? That would seem to make it look infinite, imo. And to achieve 4+GHz with nothing more than a liquid cooling setup and not a more exotic cooling solution (ie. LN2, peltier, phase change) would be impressive.

As for the TDP...who knows. I thought AMD used the ACP rating for it's thermal limits.

"Massive Headroom"?
Does that refer to the empty space inside the head of a person that goes on OCing that processor?

Dont forget to take the 750SB into account here.

Seriously now, are you suggesting, that AMD is suggesting, that these will overclock infinitly because they put an arrow on top of the bar? I am literally LOL......

I'll suggest that AMD assumes anyone reading that slide have a little more common sense than that. Well, evidently not everyone.

And finally, this is starting to sound pretty desperate, NMDante:

Now, surely you can see the lolz in that statement.

The BE's are just that, and not 1000$. Im implying that alot of people can afford a BE, and tout away they will. I just hope what Ive heard either comes true, or makes no difference. Ive heard Intel may have a few newer cpus to come out yet....possibly faster ones

don't you remember that video clip from year ago where they had this young boy demonstrate how easy it was to overclock using the overdrive application??

Good catch on Turbo, but Turbo is an application based core control, which gives a temporary overclock, than a real overclock.

Using Turbo will not overclock itself up to 3.8GHz on air, and 4+Ghz using liquid cooling, will it?
Are ACC and Overdrive all automatic overclocking on AMD systems, where you just set the frequency you want it to get to, and the CPU and motherboard will just get to that frequency?

Turbo, as I understand it, turns off cores when not needed, especially in single threaded apps. It does nothing permanent to the CPU frequency.
From Tom's:

[url=http://www.tomshardware.com/news/Turbo-Mode-Intel,6193.html][/url]

To be honest, no. I don't really trust any application based overclocking, only because it's too easy to push the clock way to far than what the CPU or motherboard can handle, and by the time you realize what happened, it might be too late to stop the damage. They seem similar to ASUS' or Gigabyte's overclocking utilities included with their motherboard drivers CD.
I'm sure there may be safeguards to prevent going "crazy" on an overclock, but if I was to overclock again, I would do it via BIOS.

i think the arrow simply means that they can pass the 4ghz barrier... more likely to hit 4.05ghz or somewhere near (with ideal conditions of coarse)

imo the main thing these chips have going for them is that they still run on the am2+ socket with ddr2 memory. seems to make it a little more worth it if their claims for stock frequencies come true.

im not getting any hopes up for overclocking until 3rd party reviews start surfacing.

Yes, there is no real damage, as long as voltage settings aren't tweaked.
The thing about this chart is that no one has really shown Deneb's overclocking range yet, and now they are claiming it's simple, with the use of just ACC and Overdrive? I don't believe it to be that simple.

Yes, when a system is overclocked incorrectly, it usually just fails to boot, but that is when adjusting clock rates or mulitpliers, not voltages. To hit extreme frequency rates, the voltage will be adjusted, and that could result in both motherboard and/or CPU damage. What would happen if Joe Schmo puts the vCore setting to 3V, because the Windows based application allowed him to adjust vCore to that limit? Again, I have never used Overdrive, so I don't know if it has voltage limits in place, but I have seen utilities that did not have limits in place.

Again, that is not just for AMD, but Intel systems as well.

It doesn't show it hitting 4GHz on air. It shows 4Ghz and above on liquid cooling.

If you don't think HKMG is a big deal, then petition AMD not to use it. If you want to know what the big deal is with HKMG, just google it's advantages. Here's an article from IBM about HKMG.
[url=http://www-03.ibm.com/press/us/en/pressrelease/23901.wss][/url]

The vcore limit in AOD is 1.45V. There is a way to allow it to let you go to 1.55V if you tweak a config setting.

Joe the Enthusiast would have to reprogram the application to allow higher than 1.55V.

That has not been proven at those speeds. As the chips go up in frequency the amount of work per clock does not increase linearly. So the only question at this point would be which architecture would scale up better?

OK, even 3.6 on air and still no HKMG. Maybe its not really needede yet, or only on bulk at those speeds?