New AMD Bulldozer Overclocks at 8.429GHz- Breaks Record!

[Bombur]

http://blogs.amd.com/play/2011/09/09/guinness/

You can find a variety of sources on this, but there's one with a video. 8.429 GHz. And on pre-release silicon, no less. The previous top 100 overclocking records had all been done on either a Pentium 4 or a Celeron based on it.

It had long been speculated that the deeper pipeline in the Bulldozer architecture should allow for higher clock speeds. With this record I guess they were right.

Tests have been shown that even with more conservative methods, the AMD Bulldozer processors, with multiplier unlocked throughout the range, appear to scale with cold. They achieved clock frequencies well above 5GHz using only air or sub-$100 water cooling solutions.

For comparison a Core i7-990X is six cores and comes in at 3.46 GHz. While the new Core i7-2600K is four cores in 3.4 GHz.

Quite an amazing feet if you ask me. what do you all think ?

[Omniyoji]

It was only a matter of time. Glad AMD was the one to accomplish the feat.

[Callisto]

I'm still a bit of a herpaderp when it comes to tech, but is there any actual practical application for even getting speeds that high, or was this just to say they got the highest recorded speed? I'm not sure of if there's a cutoff point in terms of existing software where being above a certain speed doesn't really do you any further good.

[Isiolia]

I'm still a bit of a herpaderp when it comes to tech, but is there any actual practical application for even getting speeds that high, or was this just to say they got the highest recorded speed? I'm not sure of if there's a cutoff point in terms of existing software where being above a certain speed doesn't really do you any further good.

Getting that particular CPU clocked as high as they did was more or less just for publicity. It's more or less applying insane settings, extreme cooling, and hoping that they can show it as "stable" by running some tests. It's not really practical to run that way for long. For some people it's a hobby.

Far as applications though - sure. At the extreme end, you have things that supercomputers are used for - high end simulations for example. You've got tasks like 3D rendering that can still use every scrap of computing power you toss at 'em. The more power you can pack into a (relatively) affordable desktop, the better, for certain users.
Your average user...not so much...though developers could certainly find some uses for it - IE, on the fly video transcoding.

When you hit a lot of desktop applications though, plenty of things can already run as well as they're intended and/or get bottlenecked by something other than the CPU.
A lot of what people do day-to-day on their PCs really doesn't need a high end CPU...hence the tablet and small/light laptop markets that are ramping up.

[Callisto]

Yeah that's more of what I was wondering, how much stuff would just get bottlenecked by RAM/GPU restrictions. I do a lot of video editing and while the prospect of making it go a bit faster seems nice I haven't really built up the urge to risk frying my i7-950 over it yet, lol.

[synistar]

The thought of using liquid helium, which requires one hell of a pressure chamber, is far from practical lol. Still it's awesome to push machinery to its limits like this.

[Meresgi]

The thought of using liquid helium, which requires one hell of a pressure chamber, is far from practical lol. Still it's awesome to push machinery to its limits like this.

Who doesn't have a steady supply of liquid helium on hand at their house? I know I do!

Pretty cool thing though, glad AMD got there as i'm more of a fan of them then Intel.

[hey]

No benchmarks

[Militant]

To me it looked like they ramped up to that frequency for a few seconds than turned it down? There should be a rule in the record where it has to be sustained or run a benchmark or something.

[Darus Grey]

I don't know about the world record rules, but in the OC community "Max clock" and "Stable clock" are different categories people aim for, so there's nothing really wrong with doing that since they're not claiming it's a stable clock.

[Jefe]

Eh, they need all the publicity they can get, their stock it's pretty much toilet paper right now. Both, AMD and Intel need to embrace mobile processors or their future will be horrendous.

[Don Katsu]

Eh, they need all the publicity they can get, their stock it's pretty much toilet paper right now. Both, AMD and Intel need to embrace mobile processors or their future will be horrendous.

They are both jumping on the tablet bandwagon and right now you can play any game on a shitty triple core athlon as long as your video card is up to par. Gamers account for a small number their sales.

[Kurokikaze]

Its almost over 9000! O_O

That is impressive. I'm honestly surprised it was AMD to accomplish such a task first though. Then again they did jump on hexacore first.

[SephYuyX]

First = Worst.
Second comes along and takes the good ideas and makes it better/stable/usable.

Eh, they need all the publicity they can get, their stock it's pretty much toilet paper right now. Both, AMD and Intel need to embrace mobile processors or their future will be horrendous.

What?

In a few years they will be one in the same.
By 2015 we will be at 7nm die shrinks which will have lower power consumption than current mobile CPUs, and be faster than current desktop CPUs.

Intel doesnt have to work on a mobile CPUs because theyre doing it indirectly by shrinking naturally.

By 2020 there will hopefully be mass consumed stable quantum CPUs.

[Jefe]

What? You missed the point in its entirety. ARM processors have around 70k transistors compared to a billion~ from Sandy Bridge processor. What differentiates ARM architecture and the reason why their processors are so outstanding are, its simplicity, low power consumption, and it's super cheap to produce. Intel's presence in the mobile sector is virtually non existent, well I'd say 1/10. Moore's law is irrelevant in your comment since you're under the assumption ARM will stay as-is whilst Intel/AMD continue to advance.

[Kerberoz]

The previous top 100 overclocking records had all been done on either a Pentium 4 or a Celeron based on it.

Okay, I'm probably derping, but... really? Are they going by sheer Hz or by % over stock? Haven't really been paying attention to overclocking since the days when Athlon 1700/2500s were raping everything.

How much longer until we have memristor chips?

[Isiolia]

Okay, I'm probably derping, but... really? Are they going by sheer Hz or by % over stock? Haven't really been paying attention to overclocking since the days when Athlon 1700/2500s were raping everything.

How much longer until we have memristor chips?

By sheer Hz.

Has to do with CPU architecture mostly - the P4's Netburst architecture was more readily capable of high speeds, though it wasn't able to do as much per clock cycle.

The new Athlon chips aren't quite as extreme in design as the P4s were, but share some traits that will likely result in seeing high clockspeeds on them (longer execution pipeline).

[Gryffes]

The reason clockspeed stopped raising was because processing power stopped being linked purely to Hz.

Let's see it beat a comparable sandy bridge cpu and then start fapping.