Sunday, June 16th 2013

AMD A10-6800K Cracks 8.00 GHz Mark

AMD's new A10-6800K "Richland" APU is turning out to be quite the chip for competitive overclocking. Finnish overclocker "The Stilt" cracked the 8.00 GHz mark with the chip. Skills aside, "The Stilt" relied on liquid-nitrogen cooling that kept the chip at -185°C, a core voltage of 1.992V, a base clock of 126.99 MHz (~127 MHz), and 63.0x BClk multiplier. All four cores were enabled during the feat. Other key components of the bench include an ASUS F2A85-V Pro motherboard, and 2x 4 GB dual-channel memory of an unknown make, running at 1692 MHz (DDR) Find details and validation links here.
Add your own comment

48 Comments on AMD A10-6800K Cracks 8.00 GHz Mark

#1
newtekie1
Semi-Retired Folder
by: SeventhReign
I just really dont see the point of this. It has absolutely zero real world impact. Dont even bother trying to sell the argument that it helps the developers and manufacturers create better chips in the future. It does NOT. Until there is a Heatsink/Fan or Watercooling Kit that can cool as well as LN2 *sarcasm*, these tests are nothing more than bragging rights and a waste of time.
It is the same point as working on a car and taking it to the drag strip over and over again to get it to go faster and faster. It is fun. When you finally crack the 11 second mark it only makes you want to crack the 10 second mark. A lot of people get enjoyment out of it and won't get to 10 seconds, a few do get to 10 seconds, and a few go faster than that. They enjoy it, and a few get lucky enough to make money from it. The same applies to extreme overclocking. It is a hobby. I bet there are things you do I could say are a waste of time.
Posted on Reply
#2
jmcslob
by: newtekie1
It is the same point as working on a car and taking it to the drag strip over and over again to get it to go faster and faster. It is fun. When you finally crack the 11 second mark it only makes you want to crack the 10 second mark. A lot of people get enjoyment out of it and won't get to 10 seconds, a few do get to 10 seconds, and a few go faster than that. They enjoy it, and a few get lucky enough to make money from it. The same applies to extreme overclocking. It is a hobby. I bet there are things you do I could say are a waste of time.
Just saying...
The car analogies only hold up if you only plan on using the car once...

A better analogy would be making an accurate 3D printed gun because its fun...you may only get 1 maybe 2 shots at a long distance target but its all the effort that makes it fun.
Posted on Reply
#3
The Stilt
by: SeventhReign
Dont even bother trying to sell the argument that it helps the developers and manufacturers create better chips in the future. It does NOT.
I must disagree with you.

While most of the necessary data can and will be aqquired by testing and simulations made "in-house", there are still things which cannot either be accurately simulated or tested by using the conventional methods.

While the effects of sub-zero temperatures on semiconductors are well known and can be simulated very accurately (basic physics), there are many delicate variables which still require real world testing.

The wide variety of data aqquired by the different testing methods will help the chip engineers to understand the behavior of the part even more better. Sometimes the data aqquired by the "unconvetional" testing might reveal some pitfalls or limitations which can be easily fixed in the next chip revision. These fixes can and usually will improve the 'experience' of an average consumer too ;)

Still, I agree with you on one thing: "extreme overclocking" is nowdays mostly marketing.
In my case, the scale leans more towards to the side of development.

Every single piece of additional data is valuable and all of the manufacturers want to have it as much as they can.
In the CPU business, both of the major players: Intel and AMD.

Of course it is advisable not to believe everything that you read from the internet.

by: Jorge


Stilt has some inside connections and the ability to sort thru a lot of hardware to find what works best. I doubt many Richland APUs will OC to 8 GHz. but it just shows that AMD is developing good products with a lot of potential.
I have tested three different A10-6800K specimens.
They are all from the same batch as the parts sent to the reviewers (1303SUS).
From what I have heard the retail parts which are newer production (1314 / 1318) generally seem to overclock even higher atleast on conventional cooling.

Two out of my three 6800K chips do 8GHz with both of the CUs enabled.
The other requires 1.925V and the other a bit under 2V.

The third chips does 7936MHz.

There is clearly evolution happening as with Trinity I had to go thru around 20 chips to find one that could breach the 7.9GHz mark.

Regards
Roger
Posted on Reply
#4
drdeathx
by: The Stilt
I must disagree with you.

While most of the necessary data can and will be aqquired by testing and simulations made "in-house", there are still things which cannot either be accurately simulated or tested by using the conventional methods.

While the effects of sub-zero temperatures on semiconductors are well known and can be simulated very accurately (basic physics), there are many delicate variables which still require real world testing.

The wide variety of data aqquired by the different testing methods will help the chip engineers to understand the behavior of the part even more better. Sometimes the data aqquired by the "unconvetional" testing might reveal some pitfalls or limitations which can be easily fixed in the next chip revision. These fixes can and usually will improve the 'experience' of an average consumer too ;)

Still, I agree with you on one thing: "extreme overclocking" is nowdays mostly marketing.
In my case, the scale leans more towards to the side of development.

Of course it is advisable not to believe everything that you read from the internet.



I have tested three different A10-6800K specimens.
They are all from the same batch as the parts sent to the reviewers (1303SUS).
From what I have heard the retail parts which are newer production (1314 / 1318) generally seem to overclock even higher atleast on conventional cooling.

Two out of my three 6800K chips do 8GHz with both of the CUs enabled.
The other requires 1.925V and the other a bit under 2V.

The third chips does 7936MHz.

There is clearly evolution happening as with Trinity I had to go thru around 20 chips to find one that could breach the 7.9GHz mark.

Regards
Roger
Are you Sammy?
Posted on Reply
#5
silkstone
by: SeventhReign
I just really dont see the point of this. It has absolutely zero real world impact. Dont even bother trying to sell the argument that it helps the developers and manufacturers create better chips in the future. It does NOT. Until there is a Heatsink/Fan or Watercooling Kit that can cool as well as LN2 *sarcasm*, these tests are nothing more than bragging rights and a waste of time.
Posted on Reply
#6
The Stilt
by: drdeathx
Are you Sammy?
I'm pretty certain that I am not.

Who is Sammy?
Posted on Reply
#7
cadaveca
My name is Dave
by: The Stilt
Who is Sammy?
LOL.


:laugh:

by: The Stilt
They are all from the same batch as the parts sent to the reviewers (1303SUS).
Uh...my review sample down below. Not "All reviewers" got 1303. Just sayin':

Posted on Reply
#8
drdeathx
by: The Stilt
I'm pretty certain that I am not.

Who is Sammy?
AMD renowned overclocker. I think he is from Finland. I was at one of AMD's Tech days in Austin. LOL Liquid Helium Overclock.









Posted on Reply
#9
cadaveca
My name is Dave
by: drdeathx
AMD renowned overclocker. I think he is from Finland. I was at one of AMD's Tech days in Austin
Sami Makinen = Macci.:shadedshu
Posted on Reply
#10
drdeathx
by: cadaveca
Sami Makinen = Macci.:shadedshu
Thats him.. Thanks Dave
Posted on Reply
#11
NeoXF
If the new FXs are also from the same silicon/power management revision as the Richlands (the so called, "Vishera 2.0")... I think we can expect some records from those as well...

Who knows, maybe 9GHz+ on all 8 cores at last?
Posted on Reply
#12
Tatty_One
Is that a custom bios allowing the 63x multi?
Posted on Reply
#13
seronx
by: Tatty_One
Is that a custom bios allowing the 63x multi?
The A85X chipsets have a max supported multiplier of 63x. It isn't custom it is reference.
Posted on Reply
#14
newtekie1
Semi-Retired Folder
by: jmcslob
Just saying...
The car analogies only hold up if you only plan on using the car once...

A better analogy would be making an accurate 3D printed gun because its fun...you may only get 1 maybe 2 shots at a long distance target but its all the effort that makes it fun.
Most people do plan to only use the car once, or rather a few times until something breaks and then they replace it and repeat. Pretty much what extreme overclockers do.
Posted on Reply
#15
Fourstaff
by: newtekie1
Most people do plan to only use the car once, or rather a few times until something breaks and then they replace it and repeat. Pretty much what extreme overclockers do.
I don't see land speed record cars running more than a few times either, but they are essential for research.
Posted on Reply
#16
Tintai
Wow! But why AMD? Where is Intel and 8GHz?
Posted on Reply
#17
m1dg3t
by: Fourstaff
I don't see land speed record cars running more than a few times either, but they are essential for research.
The "Big Boys" of LSR run Jet engines and have basically no active steering, how does that apply to std vehicles? Not a good analogy/comparisson.

They are only cars in the sense that they have 4 wheels and an engine :rockout:

Std PC users have a tonne more in common with top level OCers than your everyday Joe and his sport car does with a top LSR team
Posted on Reply
#18
mandis
This is phenomenal!! Well done AMD!!! :toast:

I was thinking about going with intel for this upgrade cycle but now I'll have another look at AMD's offerings. I only do Gaming and CAD on my PC anyways... :p
Posted on Reply
#19
Fourstaff
by: m1dg3t
The "Big Boys" of LSR run Jet engines and have basically no active steering, how does that apply to std vehicles? Not a good analogy/comparisson.

They are only cars in the sense that they have 4 wheels and an engine :rockout:

Std PC users have a tonne more in common with top level OCers than your everyday Joe and his sport car does with a top LSR team
Aerodynamics, materials, tuning, safety, etc. etc.
Posted on Reply
#20
Tatty_One
I don't want the IGP so am going to get an Athlon x4 760K which is also Richland, basically the 6800K without IGP just to see what she can do. Probably spend too much on this Asus mainboard he used...... seems wrong though paying twice as much for a board than the CPU that goes in it!
Posted on Reply
#21
cdawall
where the hell are my stars
by: Tatty_One
I don't want the IGP so am going to get an Athlon x4 760K which is also Richland, basically the 6800K without IGP just to see what she can do. Probably spend too much on this Asus mainboard he used...... seems wrong though paying twice as much for a board than the CPU that goes in it!
The ASUS F2A85-V Pro is only $140?
Posted on Reply
#22
eidairaman1
by: m1dg3t
The "Big Boys" of LSR run Jet engines and have basically no active steering, how does that apply to std vehicles? Not a good analogy/comparisson.

They are only cars in the sense that they have 4 wheels and an engine :rockout:

Std PC users have a tonne more in common with top level OCers than your everyday Joe and his sport car does with a top LSR team
LSR teams do have steering, just very minute, they have to control the vehicle somehow
Posted on Reply
#23
Tatty_One
by: cdawall
The ASUS F2A85-V Pro is only $140?
UK prices sadly...... £117 :cry: The CPU is £58 (just over half the price of the 6800K) which imo is a steal, unless i try my luck with a decently priced mid range board like the MSI GD65.
Posted on Reply
Add your own comment