oc make games crash?

Toggle sidebar Toggle sidebar
N

NaXos Gaming

Commendable
Jan 7, 2017
111
0
1,710
hello i realy ned help
when i oc cpu (i5-6600k) i run furmark cpu stress (oc to 4.6@1.4, stable)
when i oc gpu (Palit jetstream 980ti) i run furmark gpu stress (oc +250 core +500 mem, stable)
but if i run them both at the same time they run for about 5 min then my screen goes black for 2 sec and then i turn back on and i know its still responding when the screen is black (tested)

when ever i play games i randomly get the csgo.exe or bf4.exe has stopped working
and sometimes i even get a bsod (kernel security check failure, or something like that)
so i googled it and and found this
http://www.thewindowsclub.com/kernel-security-check-failure-windows
i did all that the guide sayed but i still get bsod and i belive it the oc that causes this
but cinebench says im fine prime95 says im fine furmark the same
same with gpu in furmark it levels at 69-70c with a oc of +250 core +500 mem
but it is stable

when i start my pc its all nomal i get the bios screen then grub bootloader (dualboot with linux) the i get to windows and see the desktop but then i run task maneger and i se a task called vmi provider host is taking all my cpu for the first 10-15 sec
any idea? just beforce i get bsod se cpu usage climp from 60 to to 99 and then bsod but i did not have task maneger open so idk but my antivirus (malwarebytes prem) seams to take a lot of cpu together with win defender

thanks in advance
 
Solution
LowlySkeleton
There is no point in running prime95 once you encounter errors. Prime95 will automatically stop worker on that core and you will see less than 100% cpu usage.

From your tests so far, we can tell that it is fairly unlikely for there to be anything wrong with the ram or memory controller (breath a sigh of relief!). Failing small FFTs in prime95, we can tell that your cpu is not stable for some reason. The most likely scenario is that there is not enough voltage being delivered to the cpu. Prime95, in smallFFTs, will attempt to force your cpu to work at full load. At full load, the voltage delivered to the cpu will drop, as many parts of the cpu are demanding more power. This voltage drop is normal and is what we call Vdroop.

Now, back...
Sort by date Sort by votes
nzalog

nzalog

Respectable
BANNED
Jan 2, 2017
541
0
2,160
If OC it causes games to crash, something in your OC is not stable. Don't matter how well it does on Prime or furmark or benchmarks. No utility uses every feature of your system in the same manner. So now use the game that crashes as your measure of stability and fix your OC.
 
Upvote 0 Downvote
N

NaXos Gaming

Commendable
Jan 7, 2017
111
0
1,710


but its also crasing running stock
and it worked before
 
Upvote 0 Downvote
LowlySkeleton

LowlySkeleton

Reputable
Aug 5, 2015
235
1
4,860
I usually would not use furmark as an actually stress test; I mainly use that for the GPU as a burnin test, to see how how it will get under heavy load.

What type of test are you using for prime95? It is better to use smallFFTs, rather than blend or largeFFTs, as this would be a good indicator for if cpu stability. If it spits out an error or BSOD after a measly 30mins, there might be be sufficient voltage for the cpu to run at that frequency. It would also be best to have something open to monitor system temps. I feel that 1.4V Vcore is pretty much as high as you safely go on that chip; it would be ill-advised to go any higher.

Testing the cpu is also only a part of a complete series of stress tests for overclocking. You should also run Memtest, as this can tell you if you have any issues with the memory controller/northbridge or ram. If you encounter errors, it may be that the memory controller isn't getting enough voltage, the ram isn't getting enough voltage, the ram timings are too tight, etc.
 
Upvote 0 Downvote
N

NaXos Gaming

Commendable
Jan 7, 2017
111
0
1,710


ok im gonna try to run memtest but i did already run windows mem dianostic?
prime 95 tried most of the settings mostly blend
om at 1.4v and 4.6 ghz tried to go to 4.7 but then it first started to show stability at 1.47v and its way over the top
sys temp msi afterburner makes a good job
gonna try smallffts when i get to my computer (writing on laptop)
memtest, prime95 smallfft, 30min, no bsod... i got it
sry for gramma fails english is not my first language
btw at 1.4v temps are a in the higher 60'es

i also got this from userbenchmark
http://www.userbenchmark.com/UserRun/3408183
its my system

if its ram or the northbride
how do i fix?

UPDATE: 1
this is memtest right?
http://i.imgur.com/G9Pb8jI.jpg

UPDATE: 2
JUST noticed that it says / 16-16-16-39?
and that xmp is on?




 
Upvote 0 Downvote
LowlySkeleton

LowlySkeleton

Reputable
Aug 5, 2015
235
1
4,860
Finding errors in Memtest will help to narrow down which part of the system is having issues. If it really is the ram or NB/memory controller, there are many things that could be done to try to resolve the errors.

The most basic thing to try would be to increase voltage to either the ram or memory controller. Though, most of the time this is not the issue, as people do not tend to overclock their ram enough that it requires that much more voltage. Most often, it would be that the memory frequency and ram timings need to be loosened (ie lower frequency and higher CL, etc).
 
Upvote 0 Downvote
N

NaXos Gaming

Commendable
Jan 7, 2017
111
0
1,710


just notised the pic says i got xmp on with i disabled?
can it be the xmp? the ram is supoosed to be at 2400 mhz
 
Upvote 0 Downvote
LowlySkeleton

LowlySkeleton

Reputable
Aug 5, 2015
235
1
4,860
XMP profile simply allows your ram to properly run at those rated settings (ie 2400MHz).

Usually, 1 pass of Memtest will suffice (even though the they suggest at least 4); the issue is unlikely to have anything to do with ram or the memory controller at this point, as there are no errors. If you had no issues running smallFFTs in prime95 and temperatures are all good, I believe that your issue might have less to do with the system stability and more with drivers, corrupt files, or something else on the software side of things.

If possible, I suggest transferring all important files to a separate drive and fresh install windows again on your boot drive. Afterwards, it would be wise to run an antivirus scan and malwarebytes, in case of any viruses or malware is hiding in the separate drive from moving those files. This method would be a bit faster than looking for which driver is working improperly or which file is corrupted.
 
Upvote 0 Downvote
N

NaXos Gaming

Commendable
Jan 7, 2017
111
0
1,710


memtest is almost done i got all 4 pass its checking my last stick of mem right now (it have been runing for 4 hours straigt
i did already run all the thing that this suggest
http://www.thewindowsclub.com/kernel-security-check-failure-windows
and that inkludes drive verifier and antivirus but i try it anyway
i got a lot of custom windows settings lige custom windows keybind in the regedit any idea how to save them?
i know what xmp is but thanks still im at the noob stage and heading to the normal stage on this forum

UPDATE:1 I WAS JUST ABOUT TO RUN WHEN THIS STROKE
bsod with stopcode
"SYSTEM_SERVICE_EXCEPTION "

any idea i tiered of this expensive piece of junk, i build myself....
UPDATE:2
google is your friend
im getting a error in prime 95 when small ffts but no crash?
this http://i.imgur.com/4MJ77rV.jpg
and memtest spit out this
http://i.imgur.com/NqV8ocI.jpg
im still running prime and getting 55% usage? oh yea 2/4 is still running

UPDATE:3 IM GONNA GO TO BED ITS 21:30 OVER HERE IN
i really dont know whats wrong gonna try motherboard monitor as prime95 suggest it if i get a hardware fail but i still have the 2/4 test running and they seams the same
 
Upvote 0 Downvote
LowlySkeleton

LowlySkeleton

Reputable
Aug 5, 2015
235
1
4,860
There is no point in running prime95 once you encounter errors. Prime95 will automatically stop worker on that core and you will see less than 100% cpu usage.

From your tests so far, we can tell that it is fairly unlikely for there to be anything wrong with the ram or memory controller (breath a sigh of relief!). Failing small FFTs in prime95, we can tell that your cpu is not stable for some reason. The most likely scenario is that there is not enough voltage being delivered to the cpu. Prime95, in smallFFTs, will attempt to force your cpu to work at full load. At full load, the voltage delivered to the cpu will drop, as many parts of the cpu are demanding more power. This voltage drop is normal and is what we call Vdroop.

Now, back to the issue on hand. Your overclock is running 4.6GHz at 1.400V. Your voltage is pretty much as high as you should go; any higher, you may need some form of exotic cooling (either a water-loop with a ton of giant radiators, phase-change, or liquid N2). Given that this is the case, your only option is to back down your overclock a bit.

Try out 4.5GHz, 4.45Ghz, 4.4Ghz, 4.35Ghz, etc. Run prime95, with the backed down overclocks, until prime95 no longer gives you errors. Wherever it is that you end up at, that should be the most stable overclock you can achieve.

For more confidence, try running prime95 for around 12-48 hours to be absolutely certain that this is INDEED the most stable overclock setting.
 
Upvote 0 Downvote
Solution
N

NaXos Gaming

Commendable
Jan 7, 2017
111
0
1,710


ok i have clocked it at 4.5@1.38 and its stable in prime95 and i also tried intel burn test coz prime would not work and then i realized something was wrong after the first bsod and now im at 4.5

UPDATE:1 what answer do u want me to pick as solution? just quote it and i pick it thanks for all the help :wahoo:
not kidding this worked
but was the solution i was scared for
but let it be that way
 
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