Overclock.net › Forums › Intel › Intel CPUs › The Intel Devil's Canyon Owners Club
New Posts  All Forums:Forum Nav:

The Intel Devil's Canyon Owners Club - Page 1975

post #19741 of 22346
Quote:
Originally Posted by GeneO View Post

What his HWINFO has listed as Cache = 1.128 is the ring or cache voltage. What it has as VRING is who knows what.

Ah okey weird....

What voltages are shown in HWinfo is that motherboard dependant?
For instant I see he has VTT, and I dont. Im guessing it is depending on the motherboard.
post #19742 of 22346
So I've been using OCCT and Prime95 for stability testing these past few days, and it seems they aren't really any better at finding instabilities than RealBench or x264, but they do it faster. 30 minutes to an hour of those and I'm pretty confident I can call it stable. But temps are an issue at these higher voltages (1.325V VID, 1.36V Vcore) so I'm not comfortable using them for longer periods of time.

So, to the issue at hand. Seeing as Prime95 is too harsh with the temps, with my H110 barely keeping the CPU under 90C (sometimes it doesn't even manage to do that - I'll probs need delidding by the end of winter), I prefer OCCT. OCCT, however, gives me errors. It doesn't downright crash, but it just throws an error alert (That dreadful quack-quack-quack sound is haunting my dreams mad.gif ) and stops the test there. It usually happens at about 3-4 minutes. Thing is, I can't narrow it down. Tried higher and lower cache speeds and voltages, tried higher and lower I/O and SA voltages, tried higher memory voltages, higher and lower VCCIN, nothing. The only thing I didn't try again was higher VCORE because I tested that a few days ago and it went for like half an hour with no problems, but I can't for the life of me remember the other settings I was using. So could anybody throw any ideas on the table?
post #19743 of 22346
Quote:
Originally Posted by QuacK View Post

Ah okey weird....

What voltages are shown in HWinfo is that motherboard dependant?
For instant I see he has VTT, and I dont. Im guessing it is depending on the motherboard.

Yes it is motherboard/chipset dependent. And things are not always labelled correctly. I think VTT is an outdated voltage and can't tell you what it means (mine is 1.112v corresponding to nothing I know of). On mine, the ring voltage is labeled "CPU Cache (VCCRING)"

.
Edited by GeneO - 2/4/16 at 5:16pm
post #19744 of 22346
Quote:
Originally Posted by tolis626 View Post

So I've been using OCCT and Prime95 for stability testing these past few days, and it seems they aren't really any better at finding instabilities than RealBench or x264, but they do it faster. 30 minutes to an hour of those and I'm pretty confident I can call it stable. But temps are an issue at these higher voltages (1.325V VID, 1.36V Vcore) so I'm not comfortable using them for longer periods of time.

So, to the issue at hand. Seeing as Prime95 is too harsh with the temps, with my H110 barely keeping the CPU under 90C (sometimes it doesn't even manage to do that - I'll probs need delidding by the end of winter), I prefer OCCT. OCCT, however, gives me errors. It doesn't downright crash, but it just throws an error alert (That dreadful quack-quack-quack sound is haunting my dreams mad.gif ) and stops the test there. It usually happens at about 3-4 minutes. Thing is, I can't narrow it down. Tried higher and lower cache speeds and voltages, tried higher and lower I/O and SA voltages, tried higher memory voltages, higher and lower VCCIN, nothing. The only thing I didn't try again was higher VCORE because I tested that a few days ago and it went for like half an hour with no problems, but I can't for the life of me remember the other settings I was using. So could anybody throw any ideas on the table?

Are you using FFT sizes of 1344 and 1344? These catch instabilities fast and don't raise the tempos to high.

Also I couldn't narrow my instability down, I was sure it was memory related, but it turned iut to be VCCIN wasn't high enough. With the 32 GB of RAM I have, I need at lease 550mv over vcore.

.
Edited by GeneO - 2/4/16 at 5:23pm
post #19745 of 22346
Quote:
Originally Posted by GeneO View Post

Are you using FFT sizes of 1344 and 1344? These catch instabilities fast and don't raise the tempos to high.

Also I couldn't narrow my instability down, I was sure it was memory related, but it turned iut to be VCCIN wasn't high enough. With the 32 GB of RAM I have, I need at lease 550mv over vcore.

.

Yup, 1344-1344 is the only configuration of Prime95 (versions 28.5 and 28.7, at least) that I've been able to run. If FFTs get any smaller boom, instant 100C. Although, I do have to ask, should I set it to run In-Place or set an amount of RAM to test, like 8GB?

Other than that, for VCCIN I've tried 1.5, 1.55, 1.6, 1.8, 1.9 and 1.95V. Lower than 1.55V and it can't stabilize my VCORE, and anything higher than 1.9V doesn't really seem to do anything, but I can't say that for sure. Maybe I'll try 2V and 2.05V, but those numbers scare me a bit as my BIOS shows them in purple. Purple usually means it's for extreme overclocking only or something. Then, over 2.05V it's red. Red usually means "Woah, back down son". tongue.gif
post #19746 of 22346
Quote:
Originally Posted by GeneO View Post

Yes it is motherboard/chipset dependent. And things are not always labelled correctly. I think VTT is an outdated voltage and can't tell you what it means (mine is 1.112v corresponding to nothing I know of). On mine, the ring voltage is labeled "CPU Cache (VCCRING)"

I think it should be named something like RAM Termination voltage, that is what its called in my bios on my Gigabyte Z97 SOC Force.
But as you can see on my screenshot there is no such value in my HWinfo..
post #19747 of 22346
Quote:
Originally Posted by tolis626 View Post

Yup, 1344-1344 is the only configuration of Prime95 (versions 28.5 and 28.7, at least) that I've been able to run.

Do yourself and your CPU a favor and use 27.9 wink.gif
post #19748 of 22346
Quote:
Originally Posted by QuacK View Post

Quote:
Originally Posted by GeneO View Post

Yes it is motherboard/chipset dependent. And things are not always labelled correctly. I think VTT is an outdated voltage and can't tell you what it means (mine is 1.112v corresponding to nothing I know of). On mine, the ring voltage is labeled "CPU Cache (VCCRING)"

I think it should be named something like RAM Termination voltage, that is what its called in my bios on my Gigabyte Z97 SOC Force.
But as you can see on my screenshot there is no such value in my HWinfo..

Yes, so??? It is not the VTT termination voltage - that would be 1/2 of the ram voltage on my MB, which is 0.8v. As far as I can see it is not related to any of the IMC or Ram voltages in my BIOS, so I think it is has nothing to do with RAM.
post #19749 of 22346
Quote:
Originally Posted by aerotracks View Post

Do yourself and your CPU a favor and use 27.9 wink.gif
I've heard quite a lot of people recommending 27.9, but I haven't gotten around to trying it. Maybe I will, although I doubt my chip will fare any better. Temps do rise and most of the time I can't do anything about it. It needs delidding, but I'm chickening out of it all the time. tongue.gif

Thanks for pointing it out though!
Quote:
Originally Posted by GeneO View Post

Yes, so??? It is not the VTT termination voltage - that would be 1/2 of the ram voltage on my MB, which is 0.8v. As far as I can see it is not related to any of the IMC or Ram voltages in my BIOS, so I think it is has nothing to do with RAM.

On my board, what you're saying is called VTTDDR and is configurable, but on auto it does set it to half of VDIMM. VTT, as reported by HWiNFO, corresponds to the digital I/O voltage, so it's IMC related. But it's most probably funky reporting more than anything else.
post #19750 of 22346
Quote:
Originally Posted by tolis626 View Post

I've heard quite a lot of people recommending 27.9, but I haven't gotten around to trying it. Maybe I will, although I doubt my chip will fare any better. Temps do rise and most of the time I can't do anything about it. It needs delidding, but I'm chickening out of it all the time. tongue.gif
Version 26.6 and 27.9 are so much cooler than the more recent versions . Give it a try and if you're still unsatisfied with temps then you can seriously consider delidding. With the easy delidding tool now it should be fool-proof but the vice only method is still as easy
Floored
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7 4790K Gigabyte Z97X Gaming 7 MSI R9 290 Gaming 4G Kingston Low Voltage Value Ram 
Hard DriveHard DriveCoolingCooling
Crucial M500 480GB Western Digital Green 3TB Noctua NH-D15 Nidec Servo Gentle Typhoons 1850rpm x 8 
OSMonitorMonitorKeyboard
Windows 8.1 Pro 64-bit Philips 240B1CB 1920 x 1200 Samsung Syncmaster 2443BW 1920 x 1200 Filco Majestouch 2 Ninja TKL 
PowerCaseMouseMouse Pad
Seasonic X750 Corsair Obsidian 750D Zowie FK-1 Barracuda Juggernaut 
AudioOther
Corsair H2100 Wireless Headset NZXT Sentry 3  
  hide details  
Reply
Floored
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7 4790K Gigabyte Z97X Gaming 7 MSI R9 290 Gaming 4G Kingston Low Voltage Value Ram 
Hard DriveHard DriveCoolingCooling
Crucial M500 480GB Western Digital Green 3TB Noctua NH-D15 Nidec Servo Gentle Typhoons 1850rpm x 8 
OSMonitorMonitorKeyboard
Windows 8.1 Pro 64-bit Philips 240B1CB 1920 x 1200 Samsung Syncmaster 2443BW 1920 x 1200 Filco Majestouch 2 Ninja TKL 
PowerCaseMouseMouse Pad
Seasonic X750 Corsair Obsidian 750D Zowie FK-1 Barracuda Juggernaut 
AudioOther
Corsair H2100 Wireless Headset NZXT Sentry 3  
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › The Intel Devil's Canyon Owners Club