Safe Voltage/Temps for 5820k

[H3x0rc!st]

I have my 5820k running at 4.4GHz atm with a Vcore of 1.3625v

Max temp running Real Bench is around 70c and seems 100% stable

I'll hopefully get the 70c down when i replace all my 120mm fans with Noctua NF-F12 PWM's

My current setup with H100i mounted to top of case:

Is 1.3625v a safe voltage to run 24/7 with my current setup?

I was hoping to get 4.4GHz with lower volts but i just get BSOD's running Real Bench.

My Bios Settings:

Vcore: 1.3625v
CPU System Agent Voltage: 1.18125
CPU Cache Voltage: 1.25
Memory @ 2688MHz - 16 16 16 39 2T
My Real Bench score is 104,147

Can i push Vcore even further to try and reach 4.6GHz or beyond?

Just want to make sure i don't damage anything, and im a bit unsure about safe 24/7 volts for these new x99 processors

Any help is appreciated!

Thanks.

Nick

 

[H3x0rc!st]

So many replies! lol

I think i'm getting the hang of overclocking on this platform now.

CPU @ 4.375GHz with 1.35v
RAM @ 3000MHz with 1.4v at 16 16 16 39 1T - Bloody love this Avexir 32GB DDR4 2400 RAM

Seems 100% stable running realbench, super PI etc for hours on end.

Also running quite a lot cooler since i fitted 5 Noctua NF-F12 PWM's

Really enjoying x99, everything is so fast compared to the i7 920 @ 4GHz which i upgraded from

CPU Temps seem to max out at around 65c under full load in real bench so i think thats pretty good

 

[Mick-1965]

Suppose it depends on how long you want to keep it. Sure Ive read that electromigration starts at ~1.3v. How fast they degrade though is anyone's guess.

Does seem high though for 4375, I'm managing the same with a tenth less (1.25). Seems stable. 8 hours real bench stress and IBT .

 

[H3x0rc!st]

1.25v for 4375mhz? Nice chip

I get bsod at below 1.35v on mine unfortunately

 

[Mick-1965]

Have you tried increasing VCCIN .
Upping that to 1.9 helped me get higher without raising the VCore too much.

 

[H3x0rc!st]

I have input voltage at 1.9v at mo. What's yours ?

 

[Mick-1965]

As above, I have it at 1.9v.

You say you get BSOD, in what ? You're not testing with Prime95 28.4 are you ?

 

[H3x0rc!st]

Nah Prime95 is bad for these chips.

I use real bench.

I get BSOD at my current speed under 1.35v. Was hoping to get 4.5GHz but i'm happy with how fast things are at the moment

 

[tipes]

I have input voltage at 1.9v at mo. What's yours ?

That's what I set mine too as well. Seems to work pretty well.

Mark

 

[misteranonimous]

I'm currently doing 4.6 GHz on 1.265V prime95 stable.

 

[H3x0rc!st]

I'm currently doing 4.6 GHz on 1.265V prime95 stable.

Very lucky

 

[Mick-1965]

I'm currently doing 4.6 GHz on 1.265V prime95 stable.

What version of prime, and how long ?

 

[smsmasters]

Try IBT with avx at maximum RAM settings and report back.

 

[C64]

overclocking the uncore can cause instability best to leave it at it's default and find a cpu vcore and cpu ratio you want first before upping uncore
you should start with VCCIN at 2.0 volts just to eliminate that and just play with vcore and cpu ratio and leave uncore cache at default ratio(not auto) and auto cache voltage, then after finding a vcore and speed you're happy with play with uncore etc

and stress test with all power saving states disabled and voltage set manually not adaptive

overclocking uncore doesn't add much gain in performance I've read how accurate that is I am not sure but theres a thread on another overclocking site which shows benchmarks not being any higher between non overclocked vs overclocked uncore and there's barely any difference, might want to set uncore at it's default ratio to get a faster lowdown on of what vcore is required for desired clock first, it might do 4.6ghz on 1.36 volts and 45-46 uncore ratio but might do 4.6 ghz on 1.28 volts with uncore at 33-43x instead of 45-46 if that makes sense

1.36 volts seems awfully high for a 4.4ghz overclock for a haswell from what I have read around those kind of voltages should be getting you 4.7-4.8 ghz.