cancel
Showing results for 
Search instead for 
Did you mean: 

high vcore + prime 95 errors @ stock speed

Loaded_Glove
Level 7
specs:
intel i7 6700k (no OC)
Asus Maximus VIII Hero (bios 2001)
Corsair Vengeance ddr4 xmp 2666 32gb (16x2)
msi GTX 1080 Gaming X Twin Frozr VI
Samsung 950 Pro M.2 NVMe 512gb
NZXT S340
Coolermaster V850
BeQuiet! Shadow Rock Slim
NZXT FZ 140 x2 intake, NZXT FN V2 120 x2 exhaust
Windows 10 Pro 64

Been having so many issues with this build and it's been rough.

Had a Corsair HX1000 psu that was loud and an i7 6700k that I thought had high vcore so I returned them.
Keep in mind the system was stable though, passing 36 hours of memtest86 v7.1 with no errors with no OC on cpu, but ram at it's xmp speed of 2666.
Prime blend also ran on this system for 24 hours with no errors and a max temp of 76c.

Turns out I was mixing up VID and Vcore and the Vcore was actually fine.

So today I went back to the micro center and replaced that cpu before realizing the mistake I made about VID and vcore and also replacing the Corsair psu for the 1 I listed in current specs at the top of this post.

The Corsair psu was also ok btw, but I just thought it was defective because the relay click sound was just a little loud, but voltages were stable in HWinfo64 during prime.

So after getting home I put the new cpu in and the Coolermaster psu and fired up Prime. In less than 10 minutes it got an error so I panicked, but googled and decided to try setting the ram back to basic speed instead of XMP.
I didn't adjust any other settings related to the cpu or ram. I fired up Prime again to see if it would still get errors with the ram at basic speed so I could possibly rule out the ram which already passed memtest. Was really hoping I just got a bad cpu.
Then things got really interesting and in just a few minutes the temp running Prime went to 100c. Stopped the test immediately and was lucky no crash or reboot or fried cpu, but it scared the crap outta me. No errors in Prime though, but the Vcore had gone up to 1.44 which is what made the temp go so high and the cpu was only at stock speed of course.
The heat sink was definitely seated right since the idle temp went right back down instantly to 26c.

What should I do now? Did I just get a bad cpu? Could this new psu be the issue? Did my ram that previously passed memtest 36 hour run just a couple days ago got damaged somehow? Could it be the motherboard failed or got ruined?

Please help. I'm not a super noob, but I'm not very advanced either and have built systems before, but only basic setup with no tweaking or OC'ing. Just buying good parts, installing Windows and using them.
28,795 Views
90 REPLIES 90

Kenrou
Level 8
1.44v VCore is way too high for a cooler like yours, adjust LLC to 5 (will stop your voltage/temps from going too high), set it to manual 1.2v if you're not overclocking and te-test. Ram should be at XMP stock also.

OC guide : https://rog.asus.com/forum/showthread.php?78542-Overclocking-The-Core-i7-6700K-To-4-6GHz-On-The-Maxi...
CPU: i7-6700k @4.7ghz
Cooler: Noctua NH-D15
GPU: Strix 980ti OC

RAM: 2x 8GB Ripjaws V 3200mhz C14
Motherboard: Maximus VIII Hero

SSD: 2x Samsung 850 EVO 250gb
HDD: 1x WD Black 4TB
PSU: EVGA SuperNOVA G2 750W Gold
CASE: Fractal Define R5 Black

OS: Windows 8.1 Pro 64bit

Yes of course it is, but I didn't set it that high. I will recap points I made in my OP to clarify. Prime errored at cpu stock speed + XMP enabled. When this happened, vcore was much lower and temps were a bit high, but under 80c as I would expect from stock cpu speed and with the cooler I have.
I rebooted, entered the bios and simply disabled XMP to see if my ram is the issue. Once I got back into Windows and ran Prime again, the Vcore was way higher even though disabling XMP is the only thing I did. It is worth noting that the VID was only 1.4 so the motherboard was basically feeding the cpu even more voltage than the cpu was requesting.
Before disabling XMP this was not occurring and the VID was also 1.4, but the board was feeding it much less than that as you would expect. It seems that disabling XMP changed more than just the ram speed. Is that normal and if so, it seems odd that simply running the ram at default 2133 would also cause the board to feed the cpu so much voltage. It definitely seems like I must have just gotten a faulty cpu. I am aware that cpu's are the least likely thing to ever be defective right out of the box, but it could happen right?

Nate152
Moderator
Hi Loaded Glove

At default settings my 6700k draws 1.44v too, you could try resetting to defaults (F5) and set the cpu core/cache voltage voltage manually to 1.20v with LLC set to 4.

Give ROG realbench a try you'll have lower temps with it, be sure to just run the stress test and not the benchmark.

Let XMP disabled for the first run.

http://rog.asus.com/rog-pro/realbench-v2-leaderboard/

Let us know if you pass the 15 minute stress test and what cpu temps are.

Goin to the clinic this morning with my mom, but I'll try when I get home.

The thing is, I've run Prime on systems in the past and I've never seen it get a single error. Also the high vcore happened this time right after I reset everything to default so 1.44 was based on default apparently. I find it odd that simply by enabling XMP the vcore was less.
Also does your 6700 draw 1.44 with stock speed at default or do you mean OC'ed, but with the auto/default voltage. No matter what and probably just for sentimental reasons because I've always seen Prime pass that I really wanna see this system pass it again like it did before I stupidly returned the other processor.

Nate152
Moderator
Your cpu is fine. 🙂

Yes my 6700k draws 1.44v at default settings. Even though it draws 1.44v at default settings I can overclock it to 4.75GHz with 1.488v game stable..

Max safe voltage is 1.52v with a max temp of 64c as stated by Intel but you can take stress testing temps to 90c that pretty much tells you your limit. Realbench should give you about 10c lower temps.

Gaming temps and everyday use 64c but I would think up to 70c would be ok.

You could find the lowest stable voltage for 4.2GHz then enable xmp or set your ram manually, if you get errors with your ram you will need to raise the CPU VCCIO voltage and the CPU System Agent voltage.

Extreme Tweaker tab

CPU VCCIO voltage 1.20v - 1.25v
CPU Systam Agent voltage 1.20v - 1.25v

You'll only need to raise these if you get errors with your ram.

Nate152 wrote:
Yes my 6700k draws 1.44v at default settings. Even though it draws 1.44v at default settings I can overclock it to 4.75GHz with 1.488v game stable..

Max safe voltage is 1.52v with a max temp of 64c as stated by Intel but you can take stress testing temps to 90c that pretty much tells you your limit. Realbench should give you about 10c lower temps.

Gaming temps and everyday use 64c but I would think up to 70c would be ok.

You could find the lowest stable voltage for 4.2GHz then enable xmp or set your ram manually, if you get errors with your ram you will need to raise the CPU VCCIO voltage and the CPU System Agent voltage.

Extreme Tweaker tab

CPU VCCIO voltage 1.20v - 1.25v
CPU Systam Agent voltage 1.20v - 1.25v

You'll only need to raise these if you get errors with your ram.


I'm sorry but your post although correct is confusing as hell, try this :

1. Max safe temp 90c (most Skylake are perfectly fine up to around 100c but a select few don't like anything over 80c) - max safe voltage 1.45v (without LN2/Dice/heavy custom loop) - max voltage 1.52v

2. My motherboard always down-volted RAM, up to 1701 i needed manual 1.36v to run XMP, might be your issue (also my CPU at stock only uses 1.34v/1.35v)

3. VCCIO+SA is USUALLY only needed at extreme ram overclocks like 3600+

4. I understand the need to advertise RealBench but you can also use Prime95 26.6/Aida64 as they will do the same job with less heat output while stressing.
CPU: i7-6700k @4.7ghz
Cooler: Noctua NH-D15
GPU: Strix 980ti OC

RAM: 2x 8GB Ripjaws V 3200mhz C14
Motherboard: Maximus VIII Hero

SSD: 2x Samsung 850 EVO 250gb
HDD: 1x WD Black 4TB
PSU: EVGA SuperNOVA G2 750W Gold
CASE: Fractal Define R5 Black

OS: Windows 8.1 Pro 64bit

Kenrou wrote:
I'm sorry but your post although correct is confusing as hell, try this :

1. Max safe temp 90c (most Skylake are perfectly fine up to around 100c but a select few don't like anything over 80c) - max safe voltage 1.45v (without LN2/Dice/heavy custom loop) - max voltage 1.52v

2. My motherboard always down-volted RAM, up to 1701 i needed manual 1.36v to run XMP, might be your issue (also my CPU at stock only uses 1.34v/1.35v)

3. VCCIO+SA is USUALLY only needed at extreme ram overclocks like 3600+

4. I understand the need to advertise RealBench but you can also use Prime95 26.6/Aida64 as they will do the same job with less heat output while stressing.


I must have been posting my last post and editing it before I saw yours. Your post is more confusing to me than his was to you I think. You explained things very briefly and used technical terms that I don't know yet. I'm very noob when it comes to all the bios settings and while I am very thankful for your reply, I wasn't really able to follow your post. Sorry, hope I didn't just irritate you, I wasn't trying to.

~edit~ Ok so I re-read your post a couple times and after you said "try this" it seems like you just explained what numbers are safe, but didn't actually give me any specific steps to try. Am I missing something?

Oh and you mention uping ram voltage, but my kit is rated to run at it's XMP speed with just 1.2v and it passed 36 hours of memtest with no errors running off just 1.2v so is it really a good idea to up the ram voltage for my particular set then?

Loaded Glove wrote:
I must have been posting my last post and editing it before I saw yours. Your post is more confusing to me than his was to you I think. You explained things very briefly and used technical terms that I don't know yet. I'm very noob when it comes to all the bios settings and while I am very thankful for your reply, I wasn't really able to follow your post. Sorry, hope I didn't just irritate you, I wasn't trying to.

~edit~ Ok so I re-read your post a couple times and after you said "try this" it seems like you just explained what numbers are safe, but didn't actually give me any specific steps to try. Am I missing something?

Oh and you mention uping ram voltage, but my kit is rated to run at it's XMP speed with just 1.2v and it passed 36 hours of memtest with no errors running off just 1.2v so is it really a good idea to up the ram voltage for my particular set then?


Follow the tabs, also if your ram is set to 1.2v XMP and gives errors try manual 1.25v. Don't change anything else you see on my pictures other then what Nate152 or i asked you until the troubleshooting is done as most either aren't needed or might damage your hardware.

CPU Voltage

59488

RAM Voltage

59489

LLC

59490
CPU: i7-6700k @4.7ghz
Cooler: Noctua NH-D15
GPU: Strix 980ti OC

RAM: 2x 8GB Ripjaws V 3200mhz C14
Motherboard: Maximus VIII Hero

SSD: 2x Samsung 850 EVO 250gb
HDD: 1x WD Black 4TB
PSU: EVGA SuperNOVA G2 750W Gold
CASE: Fractal Define R5 Black

OS: Windows 8.1 Pro 64bit

Still running memtest86 at the moment to re-check the ram and verify it is ok and it's been running almost 11 hours now with no errors. The 1st time I ran it on this new build for 36 hours with no errors. It's regular memtest86 and not memtest 86+. The version of memtest86 is the latest version which is 7.1.
So you really think my cpu is ok? Well I'm sure you have way more knowledge and experience than me, but I've never seen Prime95 get an error before and this cpu got an error in less than 10 minutes with default settings and somewhat high voltage. Doesn't an error in Prime95 at stock settings automatically mean the hardware is defective?


So far you have suggested: "you could try resetting to defaults (F5) and set the cpu core/cache voltage voltage manually to 1.20v with LLC set to 4" and " if you get errors with your ram you will need to raise the CPU VCCIO voltage and the CPU System Agent voltage"

My ram gets no errors though, but did you mean the errors in Prime?

I'm not very experienced with bios settings so let me review your advice and make sure I am understanding all your suggestions and instructions...

1) Manually set voltage to 1.2v and LLC to 4.
Where are these settings exactly in the bios?

2) raise the CPU VCCIO voltage and the CPU System Agent voltage
These settings are only if my ram gets errors in memtest right?

Do I leave speedstep enabled and if so, doesn't manually setting the voltage just lock it to that? Does this mean even all the way down to idling at 800mhz, the cpu would still get 1.2v? Is that ok or do I need to tweak any other settings in conjunction with the manual core adjustment?

Sorry for so many questions and thank you very much for your help so far. I have built some systems over the years, but never did any advanced stuff so I am basically just 1 step above being a total noob and what I want out of this experience of all these issues, is to learn and grow and maybe even get knowledgeable enough that I can help someone some day like you guys are trying to help me.