The 2016 FX overclocking results thread

[cbn]

Impressed by the results posted in this thread which included a 5.1 Ghz FX8370E @ 1.425v and a 5.0 Ghz FX8370 @ 1.475v (both using AMD's Wraith Cooler) I thought making this thread would be a good idea.

So please post your results including processor (and when you bought it), date of manufacture (if known), clockspeed and voltage, cooler, motherboard and temperature during stress testing.

EDIT: Please also mention if your CPU came with the new updated (red fan) AMD cooling solution as seen below on this FX8300. This change in coolers were introduced sometime in Feburary so the presence would indirectly indicate a relatively new manufacture Vishera:

 

[kwikgta]

Impressed by the results posted in this thread which included a 5.1 Ghz FX8320E @ 1.425v and a 5.0 Ghz FX8370 @ 1.475v (both using AMD's Wraith Cooler) I thought making this thread would be a good idea.

So please post your results including processor (and when you bought it), date of manufacture (if known), clockspeed and voltage, cooler, motherboard and temperature during stress testing.

I those are my 2 chips you are talking about: the "e" chip is a 8370e not 8320e

 

[kwikgta]

Cpu: 8370 @ 5ghz
Voltage: 1.475v
Cooler: Phanteks 120mm u-type
Motherboard: Gigabyte 990FXA-UD5
Purchased: 2-16

 

[kwikgta]

Cpu: 8370e @ 5.1ghz
Voltage: 1.425v
Cooler: AMD Wraith
Motherboard: Gigabyte GA-990FXA Gaming
Purchased: 3-16

 

[cbn]

I those are my 2 chips you are talking about: the "e" chip is a 8370e not 8320e

Thanks. I fixed the mistake.

 

[The Stilt]

Any > 4 core FX CPU running 5.0GHz stable on Wraith cooler is a lie.

 

[LTC8K6]

Well, we at least need to see screen shots showing all the cores running at those speeds.
Throttle stop would be good for that, I think.

 

[kwikgta]

Any > 4 core FX CPU running 5.0GHz stable on Wraith cooler is a lie.

Sir, do not call me a liar!

 

[LTC8K6]

Run a benchmark, and run Throttlestop. Post a screen shot.

 

[kwikgta]

Run a benchmark, and run Throttlestop. Post a screen shot.

Nah, I don't believe I will now.....

 

[zentan]

Neither you did post any in your main thread.
Pretty simple way to prove yourself right is do post the relevant shots and make the doubters believe.
Otherwise a trend might follow up of threads with unverified claims.

 

[Leyawiin]

Nah, I don't believe I will now.....

If a claim that is very hard to attain with a water cooled system is made about an air cooled system you're going to have to substantiate it. Simple as that.

 

[MrTeal]

If a claim that is very hard to attain with a water cooled system is made about an air cooled system you're going to have to substantiate it. Simple as that.

Yeah, I'm not normally one to call people out, but 5.1GHz with the Wraith cooler hitting 55C stress testing the CPU seems like a fantasy, unless "stress testing" is sitting at the Windows desktop.

 

[SlowSpyder]

I can see his FX running at those speeds with a good helping of LLC, which he said in the other thread he had set to 'extreme' so it probably jumps above that voltage under load. But based on my experience with an FX, I find it hard to believe an air cooler, especially a middle of the road cooler like the Wraith, can adequately cool the FX humming at that speed using the actual voltage it needs at that speed for stability. I am not calling anyone here a liar, sometimes you win the silicon lottery. But, it's a pretty extreme claim in my opinion, and some screen shots / benches would go a long way as far as believing it goes. But, it's the internet, so whatever.

 

[dark zero]

Pictures or is fake that information.... No FX or SB can reach 5.0 GHz + without any help other than the fan.

 

[MrTeal]

Pictures or is fake that information.... No FX or SB can reach 5.0 GHz + without any help other than the fan.

That claim seems as unlikely as kwikgta's. 5GHz produces a lot of heat, but a big tower cooler with an aftermarket fan can dissipate a lot of heat.

 

[SlowSpyder]

That claim seems as unlikely as kwikgta's. 5GHz produces a lot of heat, but a big tower cooler with an aftermarket fan can dissipate a lot of heat.

At stupid voltages and clocks I've had my CPU alone consume over 400 watts. That's a lot of heat to remove from a square of silicon the size of a postage stamp!

 

[superstition]

I can see his FX running at those speeds with a good helping of LLC, which he said in the other thread he had set to 'extreme' so it probably jumps above that voltage under load.

Even with just one module enabled you're looking at around 1.5V to get to the Windows desktop at 5.1 GHz, with stronger air cooling than a Wraith (140mm Zalman dual tower with two fans and a 200mm fan above it as exhaust).

The Stilt will have more precise numbers but I have tested an 8320E and an 8370E with a lot of settings and several different cooling setups (including a case that I cut up a lot and having everything out of the case). I don't have a high-end board but it's digital (doubled 4) and should be good enough to get the general picture. (Maybe. The board doesn't like going past 4.4 without throttling under Prime and has BIOS bugs.)

Both the 8320E and 8370E had very similar performance. Both were lapped to 3000 grit. The 8320E was used with liquid pro on air and the 8370E was used unlapped with liquid pro and then lapped with polymer TIM. With my water loop (360mm for just CPU and three vardars) there is no way to get the results he's claiming, even with the fans on max.

The ASUS Sabertooth and Crosshair boards should be more efficient but I don't see those closing the gap even with good cooling. Perhaps with chilled water, liquid metal TIM, and a loop with Delta fans they could get to the desktop at that voltage.

 

[MrTeal]

At stupid voltages and clocks I've had my CPU alone consume over 400 watts. That's a lot of heat to remove from a square of silicon the size of a postage stamp!

You could get up that high if you're trying for max power consumption, but most shouldn't require that kind of power to hit 5GHz. TPU got just over 200W at the EPS plug on their 5GHz 8350.

The Vishera dies are still pretty large, and have the benefit of being soldered to the heat spreader. I haven't tried with my 8350 (it's in my media server and I don't want to disturb it), but from a 324mm² die I've dissipated well over 400W using a Noctua U12S with a TFC1212DE. Temps get a little extreme, but that's to be expected.

 

[SlowSpyder]

Even with just one module enabled you're looking at around 1.5V to get to the Windows desktop at 5.1 GHz, with stronger air cooling than a Wraith (140mm Zalman dual tower with two fans and a 200mm fan above it as exhaust).

The Stilt will have more precise numbers but I have tested an 8320E and an 8370E with a lot of settings and several different cooling setups (including a case that I cut up a lot and having everything out of the case). I don't have a high-end board but it's digital (doubled 4) and should be good enough to get the general picture. (Maybe. The board doesn't like going past 4.4 without throttling under Prime and has BIOS bugs.)

Both the 8320E and 8370E had very similar performance. Both were lapped to 3000 grit. The 8320E was used with liquid pro on air and the 8370E was used unlapped with liquid pro and then lapped with polymer TIM. With my water loop (360mm for just CPU and three vardars) there is no way to get the results he's claiming, even with the fans on max.

The ASUS Sabertooth and Crosshair boards should be more efficient but I don't see those closing the gap even with good cooling. Perhaps with chilled water, liquid metal TIM, and a loop with Delta fans they could get to the desktop at that voltage.

I'm going by memory, but I'm fairly certain I've gotten my FX9370 to 5+GHz at sub 1.5v levels. My FX9370 is lapped, only down to 1000 grit, but this was before I lapped it. It's worth mentioning that I'm using a Fluidyne automotive radiator with 4x200mm fans in push/pull, too.

 

[kwikgta]

Alright naysayers, I have screenshots but for the life of me I can't figure out how to post them here, so I posted them at over at overclockers.com. http://www.overclockers.com/forums/showthread.php/770632-8370e-with-wraith-cooler

 

[kwikgta]

I can see his FX running at those speeds with a good helping of LLC, which he said in the other thread he had set to 'extreme' so it probably jumps above that voltage under load. But based on my experience with an FX, I find it hard to believe an air cooler, especially a middle of the road cooler like the Wraith, can adequately cool the FX humming at that speed using the actual voltage it needs at that speed for stability. I am not calling anyone here a liar, sometimes you win the silicon lottery. But, it's a pretty extreme claim in my opinion, and some screen shots / benches would go a long way as far as believing it goes. But, it's the internet, so whatever.

It is overvolting slightly with the LLC set to extreme. I never said that the wraith is providing "adequate" cooling for this overclocked chip. It is the best of the oem coolers that I've seen, and is currently cooling the "e" chip in my benching station, which I use to do preliminary testing on my chips. Now that I've tested the 8370e and two 8370's, the 8370e will be moved to my main rig to be cooled by my Phanteks PH-TC12DX 120mm. The best 8370 will be moved to the benching station and be cooled by the wraith, to be used as my backup computer.

 

[kwikgta]

Screenshots and picture of my benching rig posted here: http://www.overclockers.com/forums/s...-wraith-cooler

 

[lord_emperor]

Cpu: 8310 4.2GHz
Voltage: 1.344v
Cooler: Coolermaster Hyper 212 Evo
Motherboard: Gigabyte GA970-DS3P
Purchased: December 2014

CPU-Z

Definitely held back by the poor motherboard (4+1 QQ). Couldn't afford better at the time.

 

[kwikgta]

Cpu: 8310 4.2GHz
Voltage: 1.344v
Cooler: Coolermaster Hyper 212 Evo
Motherboard: Gigabyte GA970-DS3P
Purchased: December 2014

CPU-Z

Definitely held back by the poor motherboard (4+1 QQ). Couldn't afford better at the time.

Yes sir, a Gigabyte GA-990FX Gaming board would really help you out.