Vishera FX8350 - Overclock record

[ShintaiDK]

You wouldnt need to boot windows when you can modify multipliers and clocks from within.

 

[Idontcare]

You wouldnt need to boot windows when you can modify multipliers and clocks from within.

Oh yeah, so the actual computing load placed on those overclocked circuits is even less. It really is just a verification that the clock generator functions without crashing and not much more. Hardly a testament to the robustness of the computing circuits themselves.

Requisite car analogy - it would be like taking an F1 race car, jacking up the chasis so the tires aren't touching the ground, taking all the tires off so there is even less resistance attributable to angular momentum, putting the transmission in gear and flooring the gas just to watch the speedometer rocket upwards and claim you set a new land-speed record (all while going absolutely nowhere).

These kinds of records really are meaningless, if you aren't doing any computing with the circuits then can you really claim you've clocked those idle circuits to 7.5GHz? I don't mind the LN2 and crazy voltages, but if they aren't checking the circuits to see if they can actually do calculations at those clockspeeds then the claim itself is rather hollow.

At any rate, anyone else find it suspicious that CPUz thinks the chip is Zambezi and gives a stepping for the chip that is a bulldozer stepping (C0)?

Given that the piledriver core itself has been released long ago as trinity, one would think CPUz has long since been coded to properly recognize piledriver versus bulldozer.

 

[ShintaiDK]

Reminds me as well of IBM and their 250Ghz transistors. Just a shame again that its only a handful that can do it. And not 1-2 billion together.

Free PR to fool the unknowing.

 

[Rvenger]

Given that the piledriver core itself has been released long ago as trinity, one would think CPUz has long since been coded to properly recognize piledriver versus bulldozer.

I'm thinking its the motherboard that is lacking a Bios update since you notice everyone is popping these chips in without updating the bios. CPU microcode is not upgraded and possibly might be hindering performance a little. *Maybe*

 

[ShintaiDK]

I'm thinking its the motherboard that is lacking a Bios update since you notice everyone is popping these chips in without updating the bios. CPU microcode is not upgraded and possibly might be hindering performance a little. *Maybe*

And those hotfixes missing...and windows 8..or maybe windows 9 will unlock the true potential

 

[NTMBK]

And those hotfixes missing...and windows 8..or maybe windows 9 will unlock the true potential

He's talking about updating the motherboard so that it correctly recognises the CPU and hence CPU-Z would report it correctly. Not really sure how you're getting from there to scheduler fixes.

 

[ShintaiDK]

He's talking about updating the motherboard so that it correctly recognises the CPU and hence CPU-Z would report it correctly. Not really sure how you're getting from there to scheduler fixes.

It was his last comment about possible performance hindering.

 

[pelov]

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 it's a case of Bulldozer/Piledriver hitting its ceiling as well as the resonant clock mesh trimming down clock speeds a bit. Trinity really doesn't overclock all that well, hitting 4.4 or 4.5ghz. The TDP is incredibly impressive given the high clock speeds and 50%-of-die VLIW4, but a 200mhz overclock is abysmal.

I'm not sure how Vishera will fare, but if we're to judge by Trinity then it isn't looking very good. Llano was hitting an additional 700mhz above stock on the same immature 32nm node.

 

[Rvenger]

It was his last comment about possible performance hindering.

I said *maybe* which was actually sort of sarcasm. There is a 98% chance it aint gonna do squat

 

[Ajay]

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.

Yeah, 4 deg kelvin vs 77 deg kelvin for LN2. Crazy stuff!

 

[Ajay]

The power supply looked something like this

Seriously, what is the point of LN2 OCs.

Awesome post!

 

[skipsneeky2]

I remember the 5ghz pentium 4 stunt that is all over youtube being such a big deal and the mobile celeron hacked into a abit motherboard that got the world record for the highest percentage overclock.

Seems soon 5ghz is gonna be the norm for ocing.

 

[Olikan]

i just wanna hear someone say "IT'S OVER 9000"

 

[Ajay]

i just wanna hear someone say "it's over 9000"

it's over 9000!

 

[Idontcare]

Seems soon 5ghz is gonna be the norm for ocing.

And the consequences will never be the same!

Web surfing and farmville apps will have never ran so fast for so many people before

And we shall call it the era of "humanity unleashed 2.0". (<- read that in your best Peter Griffin, Family Guy, voiceover)

 

[Rvenger]

The power supply looked something like this

Seriously, what is the point of LN2 OCs.

Funny

 

[tweakboy]

I rather take the 2 cores its missing rather then a 8ghz 2 core. rather have 5.5Ghz 8 core.

the 8 core will eat up the 2 core 8000Mhz hmm.

 

[Rvenger]

Does anyone here have a feeling that the power consumption is going to be very similar to Bulldozer? I think its really an architecture issue at this point but we will see.

 

[guskline]

Does anyone here have a feeling that the power consumption is going to be very similar to Bulldozer? I think its really an architecture issue at this point but we will see.

Rvenger, PileDriver will be an improvement, no doubt. However, I read one comment that seemed to imply that though the resonant mesh technology would help power consumption, it might limit the extent of overclock. Time will tell. Mine is solidly at 4.6 Ghz

 

[Greenlepricon]

Rvenger, PileDriver will be an improvement, no doubt. However, I read one comment that seemed to imply that though the resonant mesh technology would help power consumption, it might limit the extent of overclock. Time will tell. Mine is solidly at 4.6 Ghz

I understand that power consumption is supposed to be better, but those are some pretty high clocks and so maybe it will balance out? I'm hoping this isn't the case but some chips do get pretty close to their max overclock at their standard voltages, so who knows. Maybe they will sacrifice their power savings in hopes of upping the performance and competing with Intel. I'm not sure how the technology will help it or to what degree, but here's to hoping they fix that problem.

 

[guskline]

Greenlepricon: Since I own a Kill-O-meter I hooked it up to rig 3 below with the 8150 OC'd to 4.6Ghz (20 x 230 maula 1.425 vcore). I'm a cooling freak so I have the rig in a CM HAF 912 ( should have got the bigger case! I have a 200mm intake fan a side 120 mm intake fan, a 2700rpm 120 mm exhaust (panaflo) and I use a Corsair H100 at performance mode to cool the CPU. When running Intel Burn test(max stress on cpu) my rig less the monitor draws as much as 429watts - that all fans running max. At idle as I'm typing this its 146 watts. Bottomline? To get the 8150 to perform well you have to OC and use some power. I think the PileDriver will improve on that but to what degree is anyone's guess. From what has been leaked so far it will improve performance @7% to as high as 15%. I think based on the architecture limitations, 7% seems more likely. The real question will be power usuage. As they say " the jury is still out".

 

[Idontcare]

Greenlepricon: Since I own a Kill-O-meter I hooked it up to rig 3 below with the 8150 OC'd to 4.6Ghz (20 x 230 maula 1.425 vcore). I'm a cooling freak so I have the rig in a CM HAF 912 ( should have got the bigger case! I have a 200mm intake fan a side 120 mm intake fan, a 2700rpm 120 mm exhaust (panaflo) and I use a Corsair H100 at performance mode to cool the CPU. When running Intel Burn test(max stress on cpu) my rig less the monitor draws as much as 429watts - that all fans running max. At idle as I'm typing this its 146 watts. Bottomline? To get the 8150 to perform well you have to OC and use some power. I think the PileDriver will improve on that but to what degree is anyone's guess. From what has been leaked so far it will improve performance @7% to as high as 15%. I think based on the architecture limitations, 7% seems more likely. The real question will be power usuage. As they say " the jury is still out".

429W for 4.6GHz!? Yowzers. What CPU temp does that result in?

 

[sequoia464]

The real question will be power usuage. As they say " the jury is still out".

This is all that I have seen on this so far, it comes from the user at [H] that won the 8350. No NDA there.

http://hardforum.com/showpost.php?p=1039237469&postcount=104

Seems light to me at 295 watts under load, any ideas?

EDIT:
EDIT2: Looks like that is at stock - 4.0, not overclocked.

 

[Rvenger]

What I mean't is the power scaling. When you OC bulldozer past 4.2ghz it seems to just increase in large increments. I hope Piledriver's power consumption scales better with clocks.

 

[guskline]

This is all that I have seen on this so far, it comes from the user at [H] that won the 8350. No NDA there.

http://hardforum.com/showpost.php?p=1039237469&postcount=104

Seems light to me at 295 watts under load, any ideas?

EDIT:
EDIT2: Looks like that is at stock - 4.0, not overclocked

Yes that's stock. OC it to the same 4.6Ghz I have then we will have a fair comparison.