Your Haswell-E max stable OC

[LiveOrDie]

I think so too.



0.016v fluctuation at every LLC setting. Look at this screen capture I just took.
View attachment 60700 This is LLC 7, also 0.016v fluctuation. What ever LLC I use I get 3 readings. Usually the lowest reading is the most prevalent, but sometimes it's the middle one. Will test now to see if different LLC has any effect on whether the lowest or middle number is used the most during stress.

Yep you will find the voltage you set in the bios to be the max but when your system on load you will see the real voltage being use and the fluctuation between the two voltages.

I think I'm going to just stick with 4.3Ghz @ 1.25v LLC6 1.85v input this runs less input that auto does and seems to be stable so far. I was trying for 4.5Ghz but @ 1.29-1.31v i was getting some odd BSOD which was some think power related any one heard of this one?

The computer has rebooted from a bugcheck. The bugcheck was: 0x000000a0 (0x000000000000000b, 0x000000032cae6000, 0x0000000000000004, 0x0000000065abc000). A dump was saved in: C:\Windows\Minidump\120414-14562-01.dmp. Report Id: 120414-14562-01.

INTERNAL_POWER_ERROR

 

[Random Murderer]

Yep you will find the voltage you set in the bios to be the max but when your system on load you will see the real voltage being use and the fluctuation between the two voltages.

I think I'm going to just stick with 4.3Ghz @ 1.25v LLC6 1.85v input this runs less input that auto does and seems to be stable so far. I was trying for 4.5Ghz but @ 1.29-1.31v i was getting some odd BSOD which was some think power related any one heard of this one?

Not sure if X99 is the same, but for A0 stop codes on X79, raising VCCSA helped me.

 

[petedread]

901 Bios available for Rampage V

 

[petedread]

4.3 @1.25v for 24/7 is what I will be happy with, and I think I will get that. My system is overclocking in a more normal manner, now, it seems. I can easily run xmp profiles, both 2666 and 3000. No SA needed.

I'm just curious about this 0.016v thing. My CPU is currently using 1.232 under load with fluctuations between 1.216, 1.232 and 1.248v. InputV fluctuates between 1.824, 1.840 and 1.856v, 1.850 set in bios. This is the case with LLC 6 and 7. All this according to HWmonitor and Aida.

 

[EarthDog]

Voltage fluctuations have always been normal. .01 is nothing actually.

Glad to see that update 'normalized' that board with the SA.

 

[LiveOrDie]

Just a heads up on the new Rampage V bios 901 it needs more vcore to be stable at the same clocks as 802 so i flashed back lol.

 

[Jstn7477]

My Gigabyte X99-UD5 WiFi finally has a new beta BIOS after three whole months, and it seems to be more stable finally, though I needed a bit of a vcore increase (my last profile didn't work and got it stuck in a reboot loop). I literally have no time to test until the weekend, but my hopes are up that I can truly OC this stupid thing, especially the memory. The BIOSes for these Gigabyte boards seemed to be the epitome of instability, no wonder they are all rated three stars on Newegg (and here I thought I was getting something better than AsRock after several years).

 

[TheHunter]

4.3 @1.25v for 24/7 is what I will be happy with, and I think I will get that. My system is overclocking in a more normal manner, now, it seems. I can easily run xmp profiles, both 2666 and 3000. No SA needed.

I'm just curious about this 0.016v thing. My CPU is currently using 1.232 under load with fluctuations between 1.216, 1.232 and 1.248v. InputV fluctuates between 1.824, 1.840 and 1.856v, 1.850 set in bios. This is the case with LLC 6 and 7. All this according to HWmonitor and Aida.

LLC affects input voltage, set to LLC7 or LLC8.. By me LLC7 overvolts input for 0.001, 1.78v >> 1.79v, dunno could because I have more VRM phases (16 in total on deluxe) and it has stabler voltage.


Cpuv well it depends, I noticed this

cpuVID 1.23v in bios it can fluctuate a little - cpuv
cpuVID 1.234v it sets all cores to 1.249v - cpuv and kinda stays fixed there..


same @ 4.7ghz

cpuVID 1.280v fluctuates cpuv
cpuVID 1.284v sets all cores to "fixed" cpuv 1.298v

 

[LiveOrDie]

So here is my 24/7 OC seems my chip isn't to hot its more of a lemon lol. 8 hour run of LinX 0.6.4 on CPU overclock then 8 hours run on CPU+CACHE+MEMORY final voltages below. Still thinking of getting a new chip to try my luck again?

 

[LiveOrDie]

Picked up a new chip but seems to be 0.001v better than my old chip so not much gain there really at lest i'm under 1.3v for 4.5Ghz this chips needs that plus 1.9v on the the input with LLC6 old chip needed 1.31v but could pass on a 1.85v input so there very close.

 

[Mydog]

Testing how high I can get a CPUz validation with the SS, temps at -55 C

http://valid.canardpc.com/8l8fy9

 

[Random Murderer]

Testing how high I can get a CPUz validation with the SS, temps at -55 C

http://valid.canardpc.com/8l8fy9

Damn, that has to be pulling near 500W by itself!
Impressive clock with all 8c/16t.

 

[Mydog]

Damn, that has to be pulling near 500W by itself!
Impressive clock with all 8c/16t.

Actually it's kind of "fake" as I don't boot in at that speed

I boot in at 5.2 GHz with Adaptive Vcore at 1.44V and then change the MP and Vcore with TurboV Core while th Power Setting is set to Balanced, then I open CPUz validation and change Power Setting to High Performance just to save the validation. Then Changing it fast back to Balanced while I run the validation on the net.

 

[Mydog]

Memory tuning

 

[TheHunter]

Picked up a new chip but seems to be 0.001v better than my old chip so not much gain there really at lest i'm under 1.3v for 4.5Ghz this chips needs that plus 1.9v on the the input with LLC6 old chip needed 1.31v but could pass on a 1.85v input so there very close.

Well higher LLC will help stabilize Input voltage, so if you set max LLC you can use lower input voltage. LLC is only for input voltage though so no harm if its running @ max.

 

[LiveOrDie]

Well higher LLC will help stabilize Input voltage, so if you set max LLC you can use lower input voltage. LLC is only for input voltage though so no harm if its running @ max.

No using a lower LLC with a higher input voltage will work out safer as under load your input voltage can spike which is why LLC is used its used to lower your input voltage underload to supply coverage for any voltage spikes.

 

[TheHunter]

Well newer Asus bios fixed them spikes, at least @ auto level..

Older bios overvolted more, e.g. 1.76v was ok v 1.284v, now I have to use 1.79v to have same stability, max overvolt 1.809v.

 

[Suferbus]

Oh man has rendering been amazing on this thing, you need to try it!

On a side note, I am also at awe with the temps in my custom loop. I cannot believe at 4.6 the temps barely ever hit 70c on a core (Prime95) and I feel I could go so much more. I really want to push this thing to the edge but I am happy enough that because of voltage difference I dropped it to 4.5ghz for 24/7 for now because the difference did not seem to be much while the power requirements increased significantly. Temps are in the mid to low 60s with Prime95 now!

What voltages are you needing for 4.6 24/7 stable. I am at 1.35v, and it is about 8 hrs stable, then something happens on Aida64. I have not bumped up vcore yet, but max temps on Aida64 during 8hrs is 70c on core 0, mid 60's on other cores. 5930k, gigabyte x99 ud5

 

[Suferbus]

Also, what is everyone setting their LLC to? I had mine set on High, but i also have a turbo and extreme setting. I have not noticed any difference in stability.

 

[xkm1948]

ASUS Sabertooth X99 1702 version of UEFI.

5820K CPU @ 1.25V 24/7 stable at 4.2GHz.

I am running 32GB of DDR4-2800 RAM. Initially pushed the CPU to 4.6GHz, the vCore was set to auto. It seems I needed 1.3V to reach 4.6GHz stable. During AIDA64 stress test two of the six cores went over 80C so I didn't push any further. Right now I am happy with just 4.2GHz of CPU. My 5820K is air cooled with Noctua D15.

 

[Vlada011]

I just vote for i7-5820K crew and now we people with i7-5820K dominate.
Before that was exactly same %.

 

[LiveOrDie]

I can't remember what my Cache voltage was my CPU can do 4.5Ghz on 1.298v so i would say my Cache would be fine on 1.24v @ 4Ghz?

 

[Vellinious]

So...has there ever been any consensus on the voltage that will start to degrade Haswell E? I run my daily clock of 4.4 @ 1.285v and I know that's fine, but I'm more wondering about benchmarking runs. I run 1.39v for 4.7.... I'd like to run 4.8, but am thinking that 1.45v (probably about what it'd need to run it good) would be too high.

Thoughts?

 

[GhostRyder]

So...has there ever been any consensus on the voltage that will start to degrade Haswell E? I run my daily clock of 4.4 @ 1.285v and I know that's fine, but I'm more wondering about benchmarking runs. I run 1.39v for 4.7.... I'd like to run 4.8, but am thinking that 1.45v (probably about what it'd need to run it good) would be too high.

Thoughts?

That's a bit much in my book for voltage as my 5930K does not go over 1.3 for 4.5ghz and from what I have been told/heard its going to put a lot of stress on it to go beyond 1.4 which could degrade the chip faster. I would recommend keeping your clocks and settings below 1.35 if I was you.

 

[Vellinious]

That's a bit much in my book for voltage as my 5930K does not go over 1.3 for 4.5ghz and from what I have been told/heard its going to put a lot of stress on it to go beyond 1.4 which could degrade the chip faster. I would recommend keeping your clocks and settings below 1.35 if I was you.

Yeah, that's kinda what I figured. I only use the benchmark clocks for short runs here and there...I just wasn't sure if anyone was using 1.4v or higher for their benchmark runs.