Vishera FX8350 - Overclock record

[MLSCrow]

7442.83 MHz with Ln2, but they're saying standard overclockers can expect 5.0-5.5GHz...I suppose that's okay. http://wccftech.com/amd-fx-8350-overclock-7442mhz-ln2-cooling/

 

[ShadowVVL]

Seams like amd breaks the record 80-100 mgz or so every few months.

Not to be an ass but seriously I wonder if they will make it to 8 ghz before bankruptcy.

 

[looncraz]

Seams like amd breaks the record 80-100 mgz or so every few months.

Not to be an ass but seriously I wonder if they will make it to 8 ghz before bankruptcy.

They broke 8GHz with Bulldozer originally, IIRC.

 

[BallaTheFeared]

Still doing it with 2 cores/threads... I thought maybe since the clocks were so low compared to bulldozer that they were breaking some sort of "eight core" record.



They were saying bulldozer would hit 5GHz on air, and 5.5GHz on water... Run away hype train, never coming back.

 

[SlowSpyder]

5.5GHz normal overclocks would probably perform pretty well. Sadly, that's likely more marketing than truth, though.

 

[Rickyyy369]

Man I hope those 5-5.5ghz numbers are true.. At those clocks you could really see BD start to be truly competitive and less of a joke.

 

[ShadowVVL]

They broke 8GHz with Bulldozer originally, IIRC.

oh my bad thanks for the correction.

Still doing it with 2 cores/threads...

yeah that's the nasty bit,if only it was all 8 cores... if only.... but one can still dream.

 

[Rvenger]

5.5ghz on 2 cores with air

 

[Idontcare]

Here's what I don't quite get - the GloFo 32nm process has had a year to mature since bulldozer chips made on the same process were hitting over 8GHz. So why is piledriver not beating the records set by bulldozer?

Is the resonant clock mesh holding it back at the high end? Or did AMD do some crazy cherry-picking of bulldozer samples to get those records but these guys doing it with piledriver aren't getting that kind of marketing team support?

 

[BallaTheFeared]

AMD's team, with trays in tow hit their mark with the help of liquid helium which is a step colder than you can get with Ln2.

 

[tulx]

Here's what I don't quite get - the GloFo 32nm process has had a year to mature since bulldozer chips made on the same process were hitting over 8GHz. So why is piledriver not beating the records set by bulldozer?

Is the resonant clock mesh holding it back at the high end? Or did AMD do some crazy cherry-picking of bulldozer samples to get those records but these guys doing it with piledriver aren't getting that kind of marketing team support?

I think the 8150 record set at the AMD marketing event was broken several times by "private" clockers, so that's unlikely. Honestly, I don't care too much about OC performance of the 8350 under liquit nitrogen as long as the overall performance increase over 8150 is considerable and it OC's well on air.

 

[AtenRa]

Here's what I don't quite get - the GloFo 32nm process has had a year to mature since bulldozer chips made on the same process were hitting over 8GHz. So why is piledriver not beating the records set by bulldozer?

Is the resonant clock mesh holding it back at the high end? Or did AMD do some crazy cherry-picking of bulldozer samples to get those records but these guys doing it with piledriver aren't getting that kind of marketing team support?

At the time of the event, out of 7 CPUs tested only 4 could go to 8GHz win LN2

 

[inf64]

The OC record is more like a testament that design can scale higher than previous ones in terms of frequency. They are running the devices at more than double the stock clock so you can say that they are pushing it a bit. There is no use for this kind of OC though,it's more for bragging rights and serves as confirmation that pipeline can sustain it (with cold enough cooling system).

As for 5-5.5Ghz "normal" OC claims, let's wait and see. Bulldozer already can do 4.7Ghz on good air/water cooling.

 

[Bman123]

How much power does it consume running at those clock speeds tho is what I want to know

 

[BD231]

The mesh actually limits max overclock potential in trade for better power consumption at realistic clock speeds.

 

[T_Yamamoto]

I want to see AMD and Intel OC their chips using air cooling instead of ln because ln does not turn me on.

 

[guskline]

These are meaningless records. Until the top AMD desktop chip can be near the top Intel chip, this circus-like sideshow is truly embarrassing.

 

[inf64]

I guess it's important for hwbot rankings.

 

[guskline]

Here's what I don't quite get - the GloFo 32nm process has had a year to mature since bulldozer chips made on the same process were hitting over 8GHz. So why is piledriver not beating the records set by bulldozer?

Is the resonant clock mesh holding it back at the high end? Or did AMD do some crazy cherry-picking of bulldozer samples to get those records but these guys doing it with piledriver aren't getting that kind of marketing team support?

To me the answer is clear. PileDriver tweeked the 8150 to what it should have been. Unfortunately, the internal turmoil in AMD appears to be affecting even new releases. I'm glad I snagged a Bulldozer for $170. The PileDrivers are a little faster, but at what price?

 

[mikeymikec]

These are meaningless records. Until the top AMD desktop chip can be near the top Intel chip, this circus like sideshow is truly embarrassing.

I don't think that was ever AMD's priority. One needs to compare processors at the prices points they typically set at. If AMD's flagship processor sells for say $200, then you need to compare it to Intel's $200 processor.

Admittedly I don't particularly care either what a cherry-picked AMD CPU did in the best possible conditions of overclocking, it has little relevance to the average person or the gamer unless AMD is going to pick up the bill for the extra electricity, bigger PSU and liquid nitrogen usage

 

[NTMBK]

How much power does it consume running at those clock speeds tho is what I want to know

The power supply looked something like this

Seriously, what is the point of LN2 OCs.

 

[guskline]

Mikeymikec: I agree with your price point comment. Perhaps, I should have stated AMD's top cpu with Intel's top mid-level chip (3770k or for that matter 3570k)

 

[inf64]

Well since FX can be a mixed bag(performance wise) AMD priced it as such . I do think that 8350 will be great perf./price CPU, we only need to see if it improves in power draw aspect (vs FX8150).

 

[Idontcare]

How much power does it consume running at those clock speeds tho is what I want to know

Consider they are only clocking one module that high, the rest are disabled, and they are not running anything stressful like LinX.

Power consumption is probably pretty low, maybe 100W. It is basically idling and at silly cold temperatures so the leakage power will be nearly zero.

 

[Idontcare]

The OC record is more like a testament that design can scale higher than previous ones in terms of frequency. They are running the devices at more than double the stock clock so you can say that they are pushing it a bit. There is no use for this kind of OC though,it's more for bragging rights and serves as confirmation that pipeline can sustain it (with cold enough cooling system).

Not really though, not the way these "records" are validated. All we really know is that the clock generator works at that clockspeed and that the circuits needed to get into windows are functioning well enough to boot the OS at that clockspeed.

To verify the chip itself is functioning as designed they need to actually run some applications that would detect if bad math was transpiring.

Intel could clock the P3 to 1.13GHz but that didn't confirm anything in terms of the viability of the microarchitecture to run at that speed, because it couldn't do it when tasked with running real applications at that speed.