High CPU Core Voltage (stock settings)

[harney]

okay so i've reflashed the bios, loaded optimised defaults and the only thing i've done is enabled the xmp profile for my ram



ok installed aida and ran it, crashed with bsod after around 30 seconds (CLOCK_WATCHDOG_TIMEOUT)

Where are you based


Also just to confirm you have GA-Z97X-SOC Force (rev. 1.0) you are flashing

GA-Z97X-SOC Force (rev. 1.0)

http://www.gigabyte.com/products/product-page.aspx?pid=4949#bios

seen peeps flash non force board this will cause issues

Also on page 7 in your manual make sure jumper bank dimm sw & pcie sw are at default

 

[billyonecan]

i'm in Hull

Yeah, that's the correct model and revision, BIOS was at F4 (cpu core 1.419v auto) before updating to F8 (1.217v auto)

 

[harney]

i'm in Hull

Yeah, that's the correct model and revision, BIOS was at F4 (cpu core 1.419v auto) before updating to F8 (1.217v auto)

So run cpuz now goto tab motherboard and see what bios ver and date your on

your not that far then if need be...

Also look on page 24 re dual bios i would select 1 bios only and make sure your booting off that one you could be flashing a new bios but using an old back up one ...

so maybe might be wise to select 1 bios only so BIOS SW 1: Main BIOS (Boot from the main BIOS)
and select
2: Single BIOS

Also note on page 17 re ram you have 2 sticks of ram if so check there in the right slots as it states re ram configuration ...

i maybe repeating if other peeps have said but good to double check

 

[billyonecan]

cpuz is showing bios version as F8 with date 09/19/2015

 

[harney]

cpuz is showing bios version as F8 with date 09/19/2015

that's good at least we know your on the right one

so check them jumpers are at default/ set correctly .... disable XMP and have every thing set at stock default... it should run at stock & show correct volts core mem ect if not there could be an hardware issue...

you could try one stick of ram at a time

might be wise to do ram test on each stick just to rule out ram issues

 

[harney]

I assume you have a usb pen

Download - Auto-installer for USB Key (Win 9x/2k/xp/7) *NEW!*

http://www.memtest.org/#downiso

 

[billyonecan]

checked the jumpers, they're in the default position (on). bios is at full defaults with xmp disabled, will run memtest now. am i checking both sticks at once, or do i need to remove one and run it, then swap with the other and run it?

core voltage is 1.224v
ram voltage is 1.500v (1333MHz)

 

[harney]

checked the jumpers, they're in the default position (on). bios is at full defaults with xmp disabled, will run memtest now

both are fine to memtest but if you get errors than we do 1 at a time

do 1 or 2 full passes will take some time if you get no red errors all good with mem

then we can get back to finding what's the issue with your system

did you check that memory is placed in the correct slots as seen on page 17 Dual Channel Memory Configuration

 

[billyonecan]

yep the memory is placed in the correct slots

memtest is currently running now, will let you know the results. do you know roughly how long 2 passes will take?

 

[harney]

yep the memory is placed in the correct slots

memtest is currently running now, will let you know the results. do you know roughly how long 2 passes will take?

8gb 30 40mins maybe

 

[billyonecan]

ok so the first pass completed without error in 20 minutes, the second pass is currently at 46% after a total time of 48 minutes, still no errors but much slower than the first pass

 

[harney]

ok so the first pass completed without error in 20 minutes, the second pass is currently at 46% after a total time of 48 minutes, still no errors but much slower than the first pass

Yes just let it complete

It should pass but you never know..

once it does pass

Next step is just try 1 stick of ram test game at stock see what happens monitor your temps ect ...

it all so might be worth flashing the 2nd backup bios to the latest also..
did have a build with a non gigabyte board another z97 a while ago with dual bios causing me issuies with volts solved by flashing both bios

On a side note
re power stepping this should sort that
https://forums.overclockers.co.uk/showthread.php?p=29239070&posted=1#post29239070

your core should not be exceeding 1.3 never mind over1.4 usually 1.20 /1.25 depending on the chip as all 4790 will vary to some degree

Also you mentioned you have move all parts into a new case ect and you think that its been running 1.4v on the core since you got it so i would double check all connections even unplug re plug in and reseat the gfx as well ...as it may not just be a core voltage where dealing with here but possibly other issues..

I assume its passed the mem test so that's the mem ticked on the check list but it may still be the culprit not having the right volts

 

[billyonecan]

yep, memtest completed 2 passes without error in 90 minutes.

after i'd swapped everything over to the new case it was still rock solid stable, played bf1 and a number of other games for a few hours without issue. it's only since the core voltage has decreased (as a result of updating in the bios) that i've had stability issues.

i managed to run aida64 for around 10 seconds, and then stop it before freezing up and noticed that one of the cores spiked to 83°C, but it quickly came down once the fans spun up. None of the core's voltage went beyond 1.237v (see image below).

 

[harney]

yep, memtest completed 2 passes without error in 90 minutes.

after i'd swapped everything over to the new case it was still rock solid stable, played bf1 and a number of other games for a few hours without issue. it's only since the core voltage has decreased (as a result of updating in the bios) that i've had stability issues.

i managed to run aida64 for around 10 seconds, and then stop it before freezing up and noticed that one of the cores spiked to 83°C, but it quickly came down once the fans spun up. None of the core's voltage went beyond 1.237v (see image below).



would it be worth removing the heatsink, cleaning the cpu and reapplying it with some fresh thermal paste?

Yes you may as well i would also re seat the cpu while your at it and check the grid area pins on the board .....

its should be stable at stock / 1.237 but what may have happened as you have been running at 1.4 at high temps for 2 years could have caused issues and now it may want more juice than stock.... just a thought ...so maybe worthplacing 1.65 on ram and increase core to tad bit more and see if your stable

So yes reseat cpu gfx check all cables may as well

and you said and then stop it before freezing up what makes you think that it would or am i missing some thing here ...

will be away for a few hours but will get back and help when i can i am sure other peeps will assist until i am back

good luck

 

[billyonecan]

should I increase voltage on the ram but leave the timings as they are for 1333mhz? My xmp sets 1.65v and increases the timings for 2133MHz.

I haven't managed to run aida for more than 20 seconds previously - it freezes up and blue screens with CLOCK_WATCHDOG_TIMEOUT

 

[harney]

should I increase voltage on the ram but leave the timings as they are for 1333mhz? My xmp sets 1.65v and increases the timings for 2133MHz.

I haven't managed to run aida for more than 20 seconds previously - it freezes up and blue screens with CLOCK_WATCHDOG_TIMEOUT

Yes keep timing as is but set memcore set to 1.65 leave your vcore and re run test and tell me the vcore if its changed see what happens

 

[billyonecan]

ok ram voltage is @ 1.65v without xmp (running at 1333Mhz). I ran Aida for 2 minutes without crashing, cpu core voltage was 1.237 on one core, and 1.236 on the other 3. CPU temps got quite high still (i presume this is to be expected while stress testing?) - 83°C peak on Cores #0 and #1, 82°C on Core #2, and 81°C on Core #3, though for the most part temps sat around mid to high 70s on all 4 cores.

I haven't reseated the cpu or heat sink yet as I don't have any thermal compound. I found some cheap old crappy stuff, but I don't want to risk it. Will order some decent stuff before I do it.

What's the next step? Should I run Aida again for longer, or is it a bad idea given the temperatures that the cores are reaching?

edit: just had another bsod, was just casually browsing at the time :|

 

[harney]

ok ram voltage is @ 1.65v without xmp (running at 1333Mhz). I ran Aida for 2 minutes without crashing, cpu core voltage was 1.237 on one core, and 1.236 on the other 3. CPU temps got quite high still (i presume this is to be expected while stress testing?) - 83°C peak on Cores #0 and #1, 82°C on Core #2, and 81°C on Core #3, though for the most part temps sat around mid to high 70s on all 4 cores.

I haven't reseated the cpu or heat sink yet as I don't have any thermal compound. I found some cheap old crappy stuff, but I don't want to risk it. Will order some decent stuff before I do it.

What's the next step? Should I run Aida again for longer, or is it a bad idea given the temperatures that the cores are reaching?

edit: just had another bsod, was just casually browsing at the time :|

Where getting there....

Temps a little high but it is stress testing ...

We need to compare another 4790k on that cooler setup you have to really know ...Have a scout on the net see what you can find the same setup running aida..cpu and cooler you have

So if its still BSOD maybe we need a fresh slate of an OS is it possible to reinstall ..then take it from there we need to know if its the os at fault here ...

tbh
I suspect we need to increase the vcore a little but lets do the os 1st

as it should all run at stock ....

If you can reinstall and you run into issues doing so re freezing or bsod then just increase the vcore a little until you complete the install i doubt you will but will see..

worse comes to worse i am based Leeds / Wakefield about area 45 50miles so it not that far ...

but lets see how we go with a clean os that is if your are able/ want too...if you do its the same old ...make sure you back up all your personal data ect

 

[billyonecan]

okay done a fresh install of windows 10.. installed hwinfo and aida, ran for around 30 seconds on stock settings before blue screening with MACHINE_CHECK_EXCEPTION (this is the first one i've seen that isn't CLOCK_WATCHDOG_TIMEOUT).

I loaded aida up again, this is what it shows after the latest bsod



this is with xmp enabled. if i disable it it's stable for longer, but inevitably freezes up

 

[billyonecan]

another strange thing i've noticed. my corsair h75 has the two radiator mounted fans connected with the y-cable supplied by corsair to the cpu_fan header. the pump is connected to SYS_FAN6 as it sits next to the cpu, however, it doesn't appear in the bios? the bios only shows system fans 1, 2, 3 and 4. i have two 140mm case fans connected to 3 & 4 which show up with their fan speeds normally. 5 and 6 don't appear in there at all.

also, if i manually set my cpu's voltage, it'll stay at that voltage (according to hwinfo and hwmon), whereas if i set it to auto it'll fluctuate based on load (~0.725v idle, 1.236/7v under a stress test). it hit 1.236v on 3 cores, and 1.237v on 1 core after a 2 minute stress test) - no crash (xmp disabled)