Saturday, October 29th 2011

AMD OC Record Broken, Still Powered by AMD FX-8150

In mid-September, earlier this year, a team of overclockers sponsored by AMD set a new Guinness Record for clock speed by a silicon processor, setting an AMD FX-8150 processor to run at a staggering 8429.38 MHz. If anything, the coveted Guinness Record feat helped cement the general notion that AMD FX processors are good at overclocking. Sadly, AMD's record didn't last long, with renowned overclocker Andre Yang breaking it with his 8461.51 MHz feat. At this point we don't know if Andre had Guinness covering his feat to he could officially break AMD's record. AMD wouldn't mind it at all, because the new record was set using an AMD FX-8150, too. Andre did it single-handed, or at least he is the only person in the "Submitted by" field on the CPU-Z Validation page.

According to the validation page, 8461.51 MHz was achieved using a base clock speed of 272.95 MHz, with 31.0X multiplier, and a brutal core voltage of 1.992V (almost 2 volts!). As with AMD's record feat, an ASUS Crosshair V Formula motherboard was used. A single 2 GB Corsair-made memory module was used doing 909.8 MHz (1818.16 MHz DDR) with timings of 9-9-9-24T. Like with AMD's feat, only two out of the FX-8150's eight cores were enabled. More details are awaited.
Add your own comment

110 Comments on AMD OC Record Broken, Still Powered by AMD FX-8150

#1
dicobalt
by: newfellow
As far I recall a 10Ghz was planned ahead on corps while they decided to go to threading.

It's sad to see what we have to deal with today, it's still a one hell of an score there.
Intel thought they could do 10GHz by the year 2005 with the 65nm Pentium 4, note the date on the article below: by Anand Lal Shimpi on 12/11/2000

http://www.anandtech.com/show/680/6
Posted on Reply
#2
lashton
hmmmm

by: NC37
Intel will reach a wall with extreme cold so no matter how cold you get it or fast you push it, they just can't technically do it as well as AMD can due to AMD engineering to handle such low temps. But outside of testing or maybe scientific environments, not a feature the regular consumer will ever use.

But this is all irrelevant. We'll eventually get new design materials in mass production and finally see clocks much much higher than any of this. Been a lot of promising research over the past decade that is being put towards this.
This is the reason why a core i7 will NEVER beat a FX8150 in an overclocking race
Posted on Reply
#3
lashton
by: happita
I wonder if the performance is hindered by not incorporating the Hi-k Metal Gate process in this generation of CPUs. I know that they are supposed to finally make the jump with Piledriver when they come out. Intel's been using HKMG I think for 2 or 3 generations now, and it's been sucessful thus far. Hurry up and get the lead out AMD!! Show Intel that you can work fast too...:rolleyes:
HKMG wont inprove bulldozer performance, what they need to do is offer each core 12 MB or level 3 cache, also get the "modules" to split 64 but insdtructions and add a Clock cycle into the FPU, this would consume more power but SMOKE the Core i7 to hell and back
Posted on Reply
#4
n-ster
Lets all hate AMD because some guy OCed AMD's chip and broke a record..

While not a big improvement, it only really hows that AMD shine in extreme OCing. It doesn't show that Bulldozer is a pile of shit or anything like that. So shut your mouths and keep you hate for the next intel vs AMD thread k?

I wonder, what is the best 24/7 STABLE OC?
Posted on Reply
#5
Dent1
by: n-ster
Lets all hate AMD because some guy OCed AMD's chip and broke a record..

While not a big improvement, it only really hows that AMD shine in extreme OCing. It doesn't show that Bulldozer is a pile of shit or anything like that. So shut your mouths and keep you hate for the next intel vs AMD thread k?

I wonder, what is the best 24/7 STABLE OC?
I agree. It's like any thread mentioning something positive about AMD people have to add their 2 cents about Bulldozer being XYZ.

Still waiting for Hustler to explain why AMD should "give it up".

Syborfical, no need to correct me. Derailed the point I was hammering home to Hustler.
Posted on Reply
#6
Super XP
This is great, I can see the new B3 revisions pushing a lot more than 8+ GHz :D
Posted on Reply
#7
Dent1
by: Super XP
This is great, I can see the new B3 revisions pushing a lot more than 8+ GHz :D
Not too fussed, would rather the B3 revision improve single threaded performance than OC'in yield.

Not sure if its worth releasing the B3 revision Q1 2012 and Piledriver also in Q1/Q2 2012 - I feel they need to bring the B3 ASAP to space out the release dates.
Posted on Reply
#8
Super XP
by: Dent1
Not too fussed, would rather the B3 revision improve single threaded performance than OC'in yield.

Not sure if its worth releasing the B3 revision Q1 2012 and Piledriver also in Q1/Q2 2012 - I feel they need to bring the B3 ASAP to space out the release dates.
Well that is what I think they plan on doing according to some rumours. FX 8170 is said to be B3 that should get released early Q1 2012, then Piledriver follows IMO with another revision perhaps the B4 or something unless AMD can achieve the performance they want with B3. Though you are correct, it would be better for them to try and increase Single Threaded Performance NOW which will give Multi-Threaded time to mature.
Posted on Reply
#9
Zubasa
by: Dent1
Not too fussed, would rather the B3 revision improve single threaded performance than OC'in yield.

Not sure if its worth releasing the B3 revision Q1 2012 and Piledriver also in Q1/Q2 2012 - I feel they need to bring the B3 ASAP to space out the release dates.
The single most important thing B3 must achieve is lower the monstrous power consumption.
Posted on Reply
#10
qubit
Overclocked quantum bit
I'll bet Asus Formula mobos are now flying off the shelves since these overclocking world records have been set. ;) Can you get any better marketing?

I wonder if other brands could also achieve such impressive overclocks with the same chip?
Posted on Reply
#11
Recus
Some guys are funny. :D You fear the trolling because of the clear fact: you have Bulldozer, but won't get OC record. Only top overclockers will get it. Deal with it. :cool:
Posted on Reply
#12
Super XP
by: Recus
Some guys are funny. :D You fear the trolling because of the clear fact: you have Bulldozer, but won't get OC record. Only top overclockers will get it. Deal with it. :cool:
Dry Ice or Liquid N anybody :D
I can care less about Overclocking so long as gaming is top notch.
Posted on Reply
#13
Recus
by: Super XP
Dry Ice or Liquid N anybody :D
I can care less about Overclocking so long as gaming is top notch.


Not worth extra 7 fps.
Posted on Reply
#14
cdawall
where the hell are my stars
Phew and here I was thinking this thread wouldn't be full of intelidiots screaming about cores disabled specific memory used and no stability testing.
Posted on Reply
#15
dicobalt
by: cdawall
Phew and here I was thinking this thread wouldn't be full of intelidiots screaming about cores disabled specific memory used and no stability testing.
FYI for everyone, all high overclocks (both Intel and AMD) on super cooling always use very little RAM and always run on 1 core or 1 module. That's just how they do it because that's what it takes to maximize stability and get those kind of clock rates.
Posted on Reply
#16
cdawall
where the hell are my stars
by: dicobalt
FYI for everyone, all high overclocks (both Intel and AMD) on super cooling always use very little RAM and always run on 1 core or 1 module. That's just how they do it because that's what it takes to maximize stability and get those kind of clock rates.
Sounds about like what I did back in the day.
Posted on Reply
#18
qubit
Overclocked quantum bit
by: dicobalt
FYI for everyone, all high overclocks (both Intel and AMD) on super cooling always use very little RAM and always run on 1 core or 1 module. That's just how they do it because that's what it takes to maximize stability and get those kind of clock rates.
Yes, indeed. What would really impress me is if something like 10GHz could be achieved easily with a fully functional chip. Since the power wall was hit around 2003/4, clock speed increases have been pants. :ohwell:

What really surprises me though, is that despite this problem, modern CPUs since the Core 2 Duo era have been able to hit 4GHz easily on air - my E8500 is at 4.11GHz with a very standard Zalman cooler, for example. Therefore, why don't they officially release CPUs rated for 4GHz when this speed is so easily achievable?
Posted on Reply
#19
Dent1
by: qubit
Therefore, why don't they officially release CPUs rated for 4GHz when this speed is so easily achievable?
We are in a day and age where organisations are trying to show consideration to the environment.

It doesn't look attractive selling a CPU with a TDP rating of 250W for the sake of giving you 4GHz out of the box.

Also, as nINJAkECIL said, the increased temps would require Intel or AMD to put bigger heatsinks in the box and increase the overal price of the CPU.
Posted on Reply
#20
nINJAkECIL
by: qubit
Yes, indeed. What would really impress me is if something like 10GHz could be achieved easily with a fully functional chip. Since the power wall was hit around 2003/4, clock speed increases have been pants. :ohwell:

What really surprises me though, is that despite this problem, modern CPUs since the Core 2 Duo era have been able to hit 4GHz easily on air - my E8500 is at 4.11GHz with a very standard Zalman cooler, for example. Therefore, why don't they officially release CPUs rated for 4GHz when this speed is so easily achievable?
Because then the TDP would be idiotically high. And what would be the temp with those tiny hsf?
That's not pretty in marketing language.

Edit:
Got ninja-ed. lol.
Posted on Reply
#21
dicobalt
by: Dent1
We are in a day and age where organisations are trying to show consideration to the environment.

It doesn't look attractive selling a CPU with a TDP rating of 250W for the sake of giving you 4GHz out of the box.

Also, as nINJAkECIL said, the increased temps would require Intel or AMD to put bigger heatsinks in the box and increase the overal price of the CPU.
Not only that but having to create a bargain basement OEM motherboard and power supply capable of delivering that power reliably isn't exactly in the realm of major OEMs like HP and Dell. It would cause product costs to go up and major OEMs are all about keeping it cheap.
Posted on Reply
#22
cdawall
where the hell are my stars
by: dicobalt
Not only that but having to create a bargain basement OEM motherboard and power supply capable of delivering that power reliably isn't exactly in the realm of major OEMs like HP and Dell. It would cause product costs to go up and major OEMs are all about keeping it cheap.
Depends the 4ghz chip wouldn't be a base model. They use plenty good boards for top end models. Last stupid Alienware I tore apart had an Asus top end board in it. Last Asus destop I tore open had a Rampage II Gene minus some heatsinks. Top end manufactures always use high quality parts in them.
Posted on Reply
#23
qubit
Overclocked quantum bit
by: Dent1
We are in a day and age where organisations are trying to show consideration to the environment.

It doesn't look attractive selling a CPU with a TDP rating of 250W for the sake of giving you 4GHz out of the box.

Also, as nINJAkECIL said, the increased temps would require Intel or AMD to put bigger heatsinks in the box and increase the overal price of the CPU.
by: nINJAkECIL
Because then the TDP would be idiotically high. And what would be the temp with those tiny hsf?
That's not pretty in marketing language.

Edit:
Got ninja-ed. lol.
But that's just it, modern CPUs are released at up to 3.8GHz, which is very close to 4GHz, so reaching it hardly takes any more power and certainly doesn't take a bigger heatsink. The standard crappy stock one that came with my E8500 was enough for it at 4GHz and my CPU is old tech now and only rated officially at 3.16GHz.

You certainly don't need anywhere near 250W for 4GHz. :laugh: Perhaps 100W or so, even less with smaller process technology. Therefore, there would be no increase in price, either.
Posted on Reply
#24
cadaveca
My name is Dave
60000 members here, and less than 50 individuals seem to have interest in this stuff. :roll:

Neat stuff...but was this done on LN2, or LHe?

Andre Yang did this? The same guy that was banned from the OC scene for helping people cheat to qualify @ OC events?

And he's still making news in OC? Nice.

NOT.


:shadedshu
Posted on Reply
#25
theubersmurf
Not to feed the trolls here, but I am more interested in how well they overclock using conventionally available cooling (and with more moderate vcore, yikes!!!) than I am in "Xtreme" overclocking. (Or whatever)

This sort of score is great for showing off your e-peen, but not much else.
Posted on Reply
Add your own comment