[SOLVED] Intel i9 7960x , Multi-core enhancement Failure, (help requested)

Toggle sidebar Toggle sidebar
H

HaizRail007

Distinguished
Mar 21, 2015
181
8
18,695
Hello, it's been awhile since I've posted anything on here, But I've hit a wall and I'm not sure how to solve the issue. So let me start, The problem is that when I turn multi-core enhancement on in bios, and then I try to run either prime 95 or handbrake, the pc then bsod's within 10 seconds of either running prime 95 or or starting to render video using handbrake.

I've tested the cpu, checked Motherboard pins and memory modules and all are fine and dandy. I was thinking it was a motherboard issue, so I swapped out the motherboard hoping that would solve the issue. but to my dismay there was no change. the first Motherboard was a asrock x299 Killer sli/ac and the second was a gigabyte x299x arours master. In addtion I checked voltages and power draw using occt I got a reading from the cpu of 280 to 430 watts. all the power supply voltages were normal according to occt. In addition, running multi-core enhancement on, in general leaves the computer running fine except prime 95 and handbrake. I did notice it blue screened once while playing cod modern warfare, with the multi-core enhancement turned on. but if I turn it off everything runs fine on the asrock motherboard. I did send back the gigabyte board back due to it not wanting to post after a software restart. and I have bought another motherboard and I'm just waiting on it to come it. but any idea's on what is causing the multi-core to crash the computer. Moreover, I've got a buddy running that gigabyte board and a 9900x, he as well gets the bsod with multi-core enabled, when running those two programs .

pc spec's

monitors

3 24 in asus 1920 x 1080

1 40 in lg 4k tv

balistic sport 2400 mhz ddr 4 memory

cpu 7960x
gpu gigabyte aorus 2080ti WBE
m.2 860 evo
1 optical drive
1 card reader
1 hotswap
aver media capture card c985
1 mechanical hhd

cooling specs

I've tried both cooling methods

Method one

cpu cooling

480mm X 60mm raid
phobia waterblock
8 Noctua NF-F12 iPPC 3000 PWM push/pull configuration
alphacool d5 pump none pwm
2 Noctua NF-F12 iPPC 3000 PWM front Intake and 1 for exhaust

for gpu loop
2 Noctua NF-A14 iPPC-3000 PWM
280mm x 60 mm raid.

Method 2

cpu cooling
noctua d15 air cooler

psu 1300 watt super nova gold rating

gpu loop carries over and so does the 2 front intake fans and 1 exhaust fan in the rear , However, I've taken the 8 fans in the top and removed 4 and turned then into 4 additional intake fans.
and speed setting is on silent except for cpu fans which are maxed.

In addition, I've looked at event viewer on my system and my buddies and when both systems suffer from the dreaded crash, event viewer log's it as power kernal 41.

consequently, if the turbo boost 3.0 is enabled without multi-core enhancement turned on I get the same crash . but if I turn on McE and turn off all turbo boosts, it works normal and no crashes. I do feel as if it's a power issue. Furthermore, if turbo boost 3.0 along with McE is enabled and the standard turbo boost is disabled, the pc doesn't crash in prime 95 , but the video signal to my 4 monitors goes out, indicating a massive power draw.Lastly, the cpu is not being overclocked and I'm trying to get it to run with in intel's specifications.
 
Last edited:
Solution
Darkbreeze
HaizRail007 said:
I can replace the 8 sticks with 2 sets of quad for 160
Click to expand...

That just puts you right back where you are now. The whole problem is that ALL sticks need to come in ONE kit, or else there are no guarantees that they are all going to "play nice" together.

The odd man out, (Or, mixed memory)


While memory modules that did not come together in a matched set that was tested by the manufacturer to be compatible, certainly CAN still work together, often it does not. Right up front I'll tell you that if you are trying to get sticks to work in the same machine together that were purchased separately, even if they are otherwise identical according to the kit or model number or if they would seem to have identical timings and...
Sort by date Sort by votes
HaizRail007 said:
Hello, it's been awhile since I've posted anything on here, But I've hit a wall and I'm not sure how to solve the issue. So let me start, The problem is that when I turn multi-core enhancement on in bios, and then I try to run either prime 95 or handbrake, the pc then bsod's within 10 seconds of either running prime 95 or or starting to render video using handbrake.

I've tested the cpu, checked Motherboard pins and memory modules and all are fine and dandy. I was thinking it was a motherboard issue, so I swapped out the motherboard hoping that would solve the issue. but to my dismay there was no change. the first Motherboard was a asrock x299 Killer sli/ac and the second was a gigabyte x299x arours master. In addtion I checked voltages and power draw using occt I got a reading from the cpu of 280 to 430 watts. all the power supply voltages were normal according to occt. In addition, running multi-core enhancement on, in general leaves the computer running fine except prime 95 and handbrake. I did notice it blue screened once while playing cod modern warfare, with the multi-core enhancement turned on. but if I turn it off everything runs fine on the asrock motherboard. I did send back the gigabyte board back due to it not wanting to post after a software restart. and I have bought another motherboard and I'm just waiting on it to come it. but any idea's on what is causing the multi-core to crash the computer. Moreover, I've got a buddy running that gigabyte board and a 9900x, he as well gets the bsod with multi-core enabled, when running those two programs .

pc spec's

monitors

3 24 in asus 1920 x 1080

1 40 in lg 4k tv

balistic sport 2400 mhz ddr 4 memory

cpu 7960x
gpu gigabyte aorus 2080ti WBE
m.2 860 evo
1 optical drive
1 card reader
1 hotswap
aver media capture card c985
1 mechanical hhd

cooling specs

I've tried both cooling methods

Method one

cpu cooling

480mm X 60mm raid
phobia waterblock
8 Noctua NF-F12 iPPC 3000 PWM push/pull configuration
alphacool d5 pump none pwm
2 Noctua NF-F12 iPPC 3000 PWM front Intake and 1 for exhaust
2 Noctua NF-A14 iPPC-3000 PWM
for gpu loop
280mm x 60 mm raid.
Method 2

noctua d15 air cooler

psu 1300 watt super nova gold rating

gpu loop carries over and so does the 2 front intake fans and 1 exhaust fan in the rear , However, I've taken the 8 fans in the top and removed 4 and turned then into 4 additional intake fans.
and speed setting is on silent except for cpu fans which are maxed.

In addition, I've looked at event viewer on my system and my buddies and when both systems suffer from the dreaded crash, event viewer log's it as power kernal 41.

consequently, if the turbo boost 3.0 is enabled without multi-core enhancement turned on I get the same crash . but if I turn on McE and turn off all turbo boosts, it works normal and no crashes. I do feel as if it's a power issue. Furthermore, if turbo boost 3.0 along with McE is enabled and the standard turbo boost is disabled, the pc doesn't crash in prime 95 , but the video signal to my 4 monitors goes out, indicating a massive power draw.Lastly, the cpu is not being overclocked and I'm trying to get it to run with in intel's specifications.
Click to expand...

To sum the problem up, when Multi-core enhancement and turbo boost 3.0 is enabled handbrake crashes the pc 10 to 20 seconds after starting to render video and prime 95 looses video signal. if Multi-core enhancement and turbo boost 3.0 are off the pc run's properly. Now my motherboard which is the asrock x299 Killer has a separate turbo boost and when that's enabled without Multi-core enhancement and turbo boost 3.0 enabled everything is fine. and if I let it run at 2.8 with Multi-core enhancement enabled everything is fine. and if I set a core clock limit to 4.0 ghz everything is fine but if I set it to 4.2 I get the crash with theses functions enabled.

I'm not sure what's going, I tried to set the load line calibration it made no difference at all. I know it's something to do with power but I'm not sure what I need to adjust
to allow the cpu to function as intel has intended.
 
Upvote 0 Downvote
Another aspect I've noticed, this may be unrelated, is the memory, as you know x299 supports 8 dim slots and according to the asrock qvl the memory I have on hand is supported for 8 single dim's or 2 sets of quad channel. Now could there be a conflict with the quad channel memory, when it tries to switch from one set to the other.

The other reason I bring the memory up is because I've ordered a asus x299 prime edition 30 and it's qvl doesn't even list the memory single quad or even dual quad for that matter.

asrock qvl

memory on hand is curcial 8 Gb ddr 4 2400 mhz model number BLS8G4D240FSA
​


https://www.asrock.com/MB/Intel/X299 Killer SLIac/index.asp#MemorySKL




asus qvl

https://dlcdnets.asus.com/pub/ASUS/...DITION_30/X299_Series_Memory_QVL_20200526.pdf

crucial BLS16G4D240FSB.16FBD(XMP) 16GB DS Micron - 16-16-16-39 1.2 ● ● ●
crucial BLS4G4D240FSB.8FBD(XMP) 4GB SS Micron - 16-16-16-39 1.2 ● ● ●
crucial BLS8G4D240FSBK.8FBD(XMP) 8GB SS Micron - 16-16-16-39 1.2 ● ● ●
crucial CT16G4DFD824A.16FB1 16GB DS Micron D9TBH 17-17-17-39 1.2 ● ● ●
crucial CT16G4DFD824A.C16FDR1 16GB DS Micron C9BGV 17-17-17-39 1.2 ● ●
crucial CT8G4DFS824A.8FB1 8GB SS Micron - 17-17-17-39 1.2 ● ● ●


Wondering if thec qvl is wrong about the memory I have supporting 8 dim sticks and causing a crash.
 
Upvote 0 Downvote
my buddy had the same thing happen his setup running prime 95 and handbrake within the same scenario's the difference's between our systems Includes memory, cpu, storage devices. same brand memory but his is 2666 128 Gb not sure of the model and his cpu is a 9900x.
 
Upvote 0 Downvote
This is odd, I thought it could be the psu but when I've checked it, when doing stress testing it's preformed as it should. Moreover, if I do a memtest, with the functions enabled I hit 4.4 Ghz on the cpu I have been viewing occt and it has stated that my cpu was using 430 watts, and from what I've researched, some have said that it's not out of the realm of possibility for the cpu to use over 600 watts. I was looking at the breakdown of my gpu etc and occt stated that my gpu was using 300 watts per channel, I may have that one wrong. if that's right that's 1,010 and that doesn't include the motherboard power draw.
 
Upvote 0 Downvote
The most I'd expect your system to possibly pull would be about 638w, so a 750w power supply, if it is a good quality model, should be plenty. What is the exact model of your Supernova Gold PSU. Is it a G1, G2, G3, etc?

Your CPU can only pull about 163w at full load with the stock configuration.

https://www.anandtech.com/show/11839/intel-core-i9-7980xe-and-core-i9-7960x-review/14

So even if it was overclocked and we said let's increase that to 250w, a 750w PSU would still cover the roughly 725w maximum theoretical power draw, that it certainly would never come near in real life.

The 2080 ti cannot EVER pull more than 375w total, at full bore, because the two 8 pin PCIe auxiliary power connectors will only supply 150w each and the slot power is only 75w, for a total of 375w.

So, IF we SAY it is pulling 250w for the CPU, 375w for the graphics card and being very generous to the side of caution we say 100w for the rest of the system, which there is no way in hell it could ever need, that's 725w, max.

In reality if you connected a kill-a-watt to the wall socket and pulled full power from this system, it is a certainty that you would not be likely to see anything beyond 550w, MAYYYYBE as much as 600-650w if you were running Prime95 Small FFT, Furmark and Realbench all at the same time.
 
Upvote 0 Downvote
Hmm...
A)You know MCE is essentially an auto-overclock, right?
I noticed later in your initial post that you're not trying to overclock, but that's exactly what MCE is at the core... and since you're trying to overclock, you also need to remove or raise certain power limits, or it will fail.
For example: My 7820X's 4.5ghz OC would fail in Cinebench R20 even though I have the power limit and power time duration maxed out. What I did to fix it: Raise the current limit.
My mobo has 100%, 140%, and 200% settings - I don't know why the big gap between 140 and 200, but that's how it is - anyway, 140% wasn't enough.
I'm sure your 7960X draws a good deal more power and current than my 7820X does, so if you left the power and current limits on default...
But if you're not looking to overclock, you should just leave MCE alone, because that's what it is.

B)You're running 8 dimms of ram, right? Does it still crash if half of them are removed?
It wouldn't be unusual if all 8 dimms need extra Dram Channel or SOC voltage to stabilize...

C)It might even be a combined issue of both the above... or it could be completely unrelated.
 
Upvote 0 Downvote
Phaaze88 said:
Hmm...
A)You know MCE is essentially an auto-overclock, right?
I noticed later in your initial post that you're not trying to overclock, but that's exactly what MCE is at the core... and since you're trying to overclock, you also need to remove or raise certain power limits, or it will fail.
For example: My 7820X's 4.5ghz OC would fail in Cinebench R20 even though I have the power limit and power time duration maxed out. What I did to fix it: Raise the current limit.
My mobo has 100%, 140%, and 200% settings - I don't know why the big gap between 140 and 200, but that's how it is - anyway, 140% wasn't enough.
I'm sure your 7960X draws a good deal more power and current than my 7820X does, so if you left the power and current limits on default...
But if you're not looking to overclock, you should just leave MCE alone, because that's what it is.

B)You're running 8 dimms of ram, right? Does it still crash if half of them are removed?
It wouldn't be unusual if all 8 dimms need extra Dram Channel or SOC voltage to stabilize...

C)It might even be a combined issue of both the above... or it could be completely unrelated.
Click to expand...
I know that mce run's all core to max clock which is 4.2ghz and with turbo boost 3.0 4.4ghz, the reason for the change, is because I've been interested in bench mark testing for scores and seeing what I can pull out of it for future reference. The other thing is asrock uses load line caliabration 1 -5 and I've went has high as level 4 and no change. I've disabled thermal throttling the c states and I set a fixed voltage to 2.5v's for the cpu. This is about as far as my knowledge of voltages go. I can say when I set a overclock to either 4100 mhz or 4000 mhz it's fine. I'm sure it's something with the power I'm just not sure what I need to change for it to be effective.

it didn't occur to me to pull out 4 of the 8 sticks. but I will do that.
 
Last edited:
Upvote 0 Downvote
HaizRail007 said:
I've disabled thermal throttling
Click to expand...
No! Turn that back on, please.

HaizRail007 said:
I set a fixed voltage to 2.5v
Click to expand...
2.5v Input Voltage?! a;lfjadskl;fjasdklfjklj
[I'm aware that Skylake X has 2 voltage settings: Vcore and VCCIN, which is sometimes referred to Input Voltage.]
I have to wonder just how hot you made the VRMs run with 2.5v VCCIN... did you also liquid cool the VRMs?
2.0v-2.1v VCCIN should've been enough.

Between disabling thermal throttling and running 2.5v VCCIN, how do we supposed to know whether or not you damaged anything?
 
Upvote 0 Downvote
yes it's under liquid cooling, vrm's reached 55 C, and when you select the voltage under the cpu, it gives you 3 choices auto, fixed and manual. The fixed option automatically set's it to a unchangeable 2.5v and thats what the last test was set too a few days ago. in addition, believe it or not that is the default motherboard setting.

I did change it to auto just now after pulling though 4 sticks. ran prime 95 for 40 seconds or so before the screens went black. hand brake took 1 minute, so some improvement.
 
Upvote 0 Downvote
I'm not sure how I'd change the draw limit, on this asrock board every manufacturer is different, load line calibration is set to level 2 by default and I changed it to level 3. I don't want to give it to much power. because if you fry the chip it's a bad day all the way around. Probably a good thing I set it to auto instead of fixed.
 
Upvote 0 Downvote
Then you may as well leave MCE alone. It is an overclock, and that alone requires that you raise or remove certain limiters that the OC doesn't touch by default.

You're not having any serious issues with your system with MCE disabled, and it doesn't appear to be hitting any thermal limits... so that leaves power/current.
 
Upvote 0 Downvote
It's crashes with TBM 3.0? Huh, that shouldn't trigger any power/current limits.
You got me thinking it's the mobo, but you've already tried a different one with no change.


Hold on...
"curcial 4Gb X 8 sticks ddr 4 2400 mhz model number BLS8G4D240FSA" <- From your 9th post
I looked this up, and I get linked to a single 8GB stick, not an actual 8 x 8GB kit.
Is that really the model number of your ram? 8 of this is not the same as this.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom