Monitors suddenly goes black

[Naweo]

The cpu and/or/andor ram are probably not stable in the stock configuration.

Yes could be. It feels like an overclock crash to be honest. What's your suggestion for a go-to walkthrough for a clock fix?

[Leslieann]

My PC under load would consume probably 500+ W alone, and my two monitors I have plugged in under the same 900W UPS has a max consumption of 400w. That alone is 900 watt.

I don't think the UPS delivers enough power?

When under normal conditions the UPS is simply passing AC current straight through, it shouldn't be current limited in the same as when running on battery.

Also while under load we rarely ever completely load a PC 100% except under benchmark/synthetic conditions, it's just not something that happens. Even if you try, odds of the load matching perfectly with your components without one or two bottlenecks somewhere are insanely rare. There's too many combinations of hardware and software for it to happen.

Also component ratings are considered worst case, not best case and is almost always have a large buffer. It's best to plan for that much power use, but the idea is to not need it.

[tp4tissue]

The cpu and/or/andor ram are probably not stable in the stock configuration.

Yes could be. It feels like an overclock crash to be honest. What's your suggestion for a go-to walkthrough for a clock fix?

google the chipset name, and overclock guide.

Set core voltage manually, set ram timings manually <you don't need all of them just the primary timings>

Set cpu ratio to fixed, set load line calibration to medium. 

Do some testing with prime 95,  DISABLE avx , by inserting    CpuSupportsAVX=0  into local.txt in the prime95 folder.

[Naweo]

Should be able to do that.

Is it normal for stock clocks on ram/cpus to not be stable?

[tp4tissue]

Should be able to do that.

Is it normal for stock clocks on ram/cpus to not be stable?

It's not uncommon. 

Nothing's normal in the real world.  we try to make all bricks the same but the reality is they differ, and a slight deviation could make it not fit.

Default motherboard settings have also gotten extremely convoluted. 

[Naweo]

I got asrock z390

Will try 1.28v and 49 OC on my 9900k for cpu.


Any settings I missed?

For my 64 gb ddr4 2133 ram, I will find a guide. Any settings I need there?

[Naweo]

I will 24h test:

4.9 ghz at 1.3v with XMP 2.0 memory profile over nights for 24h. If stable, then I assumr my black screens are different issue

I haf 10700k, not 9900k.

So far did a 30m test with max temp of 76, 4.9 ghz on cpu, xmp 2 profile. Could not find intel C state to disable in asrock steel legend z390. 1.3v on CPU.

I see screens individually going black sometimes but no freezing.

I had my power circuit safety switch off, but might be cpincidence.

I will try to run prime 95 for at least a toral of 24h stable over a week or so and see what happens.

This is a screenshot of about 15 minutes of testing:

https://i.gyazo.com/284911ec3b9369c768b61c2ea054aadb.png

[Naweo]

Should be able to do that.

Is it normal for stock clocks on ram/cpus to not be stable?

It's not uncommon. 

Nothing's normal in the real world.  we try to make all bricks the same but the reality is they differ, and a slight deviation could make it not fit.

Default motherboard settings have also gotten extremely convoluted. 

Anything I should tweak or test? I primarily just want to find stability with the OC, not performance.

[tp4tissue]

I'm not familiar with specifics of 10700k,  but, the all core turbo is not fixed in your screenshot. it's throttling.  Could be the avx divider, could be something else.

So whatever its doing, you're still testing a floating configuration,   and the problem with THIS on the modern platform is that the transient voltages cause crashes and black screens and all sorts of gremlins.

There's alot more homework required to doing the manual tuning / testing.  Think 10+ hours of reading notes, guides, voltage monitoring/ manipulation,     It's not a set it and run prime 95 situation.

The workflow should be to first FIND the minimum stable voltage for your Peak-configuration, full all core turbo.

The performance OC may not matter to you, but THIS PROCESS of finding it, will give you incite into the stability and inner workings of your setup. That comprehension is true system awareness, and the only way to know with any certainty if the system is working as intended.     It's not an iphone.

There are no less than 40+ settings to manually Research, Understand and configure.  You have to know exactly what they do and how they interact.


Remove all your extra stuff , you need 1 monitor, keyboard mouse, that's it.

Tp4 has personally spent 100s of hours actively overclocking 60+ computers over the years.  This does not count how long prime 95 runs.   It's Tp4 in sweatpants, squinting at voltage dip graphs, reading 1000s of posts on overclocking forums.

[yui]

When i have random issues 1st thing i try is usually switching to a known good psu, i am lucky enough to have a few on hand, then if it still fail i will try a known good GPU, i have an old radeon pro for that purpose or if your cpu have an igpu it would also work, then either if i have a spare motherboard or cpu i try those, or i would underclock the cpu. i am saying that thinking that you already tried everything software and checked that your ram is properly seated, windows is rather bad at staying stable, so installing windows on an other drive and seeing if that fixes the instability could also help.
i come from a place were i have spare parts for most my pc because they are rather old and so parts were cheap or free at one point, so it helps. although to me your symptoms sounds like corrupted windows/driver install or dirty power. the other thing i saw with freezing after a long time is a mac mini G4 with a cracked chipset the cooler it is the longer it last before freezing, but that was one out of maybe one hundred or so pc that i fixed or at least tried to.
just speaking of heat, if your gpu is overheating it could also freeze the pc as it shuts itself off to save itself from burning, although that is more an AMD thing than nVidia, usually nVidia sets those securities way too high and so are never triggered.

[Naweo]

Yes I understand. I can be patient, and I am thankful you check the post all and often. I have been at this problem for months now and no reason to not try an OC solution. I can spend 20-30 minutes a day and tinker with settings ETC and just test performances overnight.

I notice there is a fluctuation in my CPU frequency. You say I need to enable turbo everywhere on my Asus z390 Steel Legend?

I'm not familiar with specifics of 10700k,  but, the all core turbo is not fixed in your screenshot. it's throttling.  Could be the avx divider, could be something else.

So whatever its doing, you're still testing a floating configuration,   and the problem with THIS on the modern platform is that the transient voltages cause crashes and black screens and all sorts of gremlins.

There's alot more homework required to doing the manual tuning / testing.  Think 10+ hours of reading notes, guides, voltage monitoring/ manipulation,     It's not a set it and run prime 95 situation.

The workflow should be to first FIND the minimum stable voltage for your Peak-configuration, full all core turbo.

The performance OC may not matter to you, but THIS PROCESS of finding it, will give you incite into the stability and inner workings of your setup. That comprehension is true system awareness, and the only way to know with any certainty if the system is working as intended.     It's not an iphone.

There are no less than 40+ settings to manually Research, Understand and configure.  You have to know exactly what they do and how they interact.


Remove all your extra stuff , you need 1 monitor, keyboard mouse, that's it.

Tp4 has personally spent 100s of hours actively overclocking 60+ computers over the years.  This does not count how long prime 95 runs.   It's Tp4 in sweatpants, squinting at voltage dip graphs, reading 1000s of posts on overclocking forums.

[Naweo]

When i have random issues 1st thing i try is usually switching to a known good psu, i am lucky enough to have a few on hand, then if it still fail i will try a known good GPU, i have an old radeon pro for that purpose or if your cpu have an igpu it would also work, then either if i have a spare motherboard or cpu i try those, or i would underclock the cpu. i am saying that thinking that you already tried everything software and checked that your ram is properly seated, windows is rather bad at staying stable, so installing windows on an other drive and seeing if that fixes the instability could also help.
i come from a place were i have spare parts for most my pc because they are rather old and so parts were cheap or free at one point, so it helps. although to me your symptoms sounds like corrupted windows/driver install or dirty power. the other thing i saw with freezing after a long time is a mac mini G4 with a cracked chipset the cooler it is the longer it last before freezing, but that was one out of maybe one hundred or so pc that i fixed or at least tried to.
just speaking of heat, if your gpu is overheating it could also freeze the pc as it shuts itself off to save itself from burning, although that is more an AMD thing than nVidia, usually nVidia sets those securities way too high and so are never triggered.

GPU definitely not overheating, the crashes happens often at 40 temp on GPU.

I will just slowly try everything until I eventually have to install windows on a new system and/or change parts. My PSU should be pretty good but I can try to exchange my PSU first.

Unfortuantely I am dependent on the monitor setup for work, so it either works or flops. The issue I have not, as long as it doesn't worsen, I can definitely live with.

[yui]

what i was saying about windows is that you should be able to install a 2nd windows on the same system, at least it was possible with 7, if the fresh windows does not crash, it is your old windows, if it does crash then the problem is somewhere else.

[Naweo]

^ Yes I see. Honestly I am trying the overclock solution and diverting the power to different outlets and two UPS (Even if one UPS is a waste of money, my brother can use it for power outages.)

If that does not work, I will just live with the issue until it gets worse I think. Will get back to you when needed.

[yui]

i was saying underclock, i had luck with that on my AMD FX8150 that had a very hot and hard life, when it became unstable to run it at stock clock running it under the stock speeds made it stable again, overvolting at stock clock could do the same but also could make things worse if the motherboard can't handle the power.
sorry if it is not clear, but your issues can actually point to pretty much any of the main component failing, it is why it is so hard to find the issue, just do not throw the UPS away unless you are very sure it is it those things often contain lead battery and so are rather bad for the environment.
and to be honest with you i do not think it will get worse, or if it does it would be very sudden if it is hardware

[Naweo]

i was saying underclock, i had luck with that on my AMD FX8150 that had a very hot and hard life, when it became unstable to run it at stock clock running it under the stock speeds made it stable again, overvolting at stock clock could do the same but also could make things worse if the motherboard can't handle the power.
sorry if it is not clear, but your issues can actually point to pretty much any of the main component failing, it is why it is so hard to find the issue, just do not throw the UPS away unless you are very sure it is it those things often contain lead battery and so are rather bad for the environment.
and to be honest with you i do not think it will get worse, or if it does it would be very sudden if it is hardware

Thank you. I will try to make a life of it.

[Naweo]

I'm not familiar with specifics of 10700k,  but, the all core turbo is not fixed in your screenshot. it's throttling.  Could be the avx divider, could be something else.

So whatever its doing, you're still testing a floating configuration,   and the problem with THIS on the modern platform is that the transient voltages cause crashes and black screens and all sorts of gremlins.

There's alot more homework required to doing the manual tuning / testing.  Think 10+ hours of reading notes, guides, voltage monitoring/ manipulation,     It's not a set it and run prime 95 situation.

The workflow should be to first FIND the minimum stable voltage for your Peak-configuration, full all core turbo.

The performance OC may not matter to you, but THIS PROCESS of finding it, will give you incite into the stability and inner workings of your setup. That comprehension is true system awareness, and the only way to know with any certainty if the system is working as intended.     It's not an iphone.

There are no less than 40+ settings to manually Research, Understand and configure.  You have to know exactly what they do and how they interact.


Remove all your extra stuff , you need 1 monitor, keyboard mouse, that's it.

Tp4 has personally spent 100s of hours actively overclocking 60+ computers over the years.  This does not count how long prime 95 runs.   It's Tp4 in sweatpants, squinting at voltage dip graphs, reading 1000s of posts on overclocking forums.

I am still managing to get throttled results, it seems. Currently 1.29v at 4700 mhz. I looked every for an AVX setting and intel-C step, but only managed to find something else. I will show a picture of the menu.

https://i.gyazo.com/4c998b9a10e5614b015219799a915b61.png

[tp4tissue]

did you set CpuSupportsAVX=0 in local.txt ?

[Naweo]

did you set CpuSupportsAVX=0 in local.txt ?

I swore I did... I will run an 8h test overnight. gotta work now.

Talk to you soon!

Thanks

[tp4tissue]

Lots of settings to play with.

[Naweo]

Lots of settings to play with.

One step at a time. And we  will get there.

[tp4tissue]

Lots of settings to play with.

One step at a time. And we  will get there.

Double check ur not thermal throttling,   if it's doing that, maybe run the test vented open chassis.

[Naweo]

Lots of settings to play with.

One step at a time. And we  will get there.

Double check ur not thermal throttling,   if it's doing that, maybe run the test vented open chassis.

I have trouble getting the PC to run at 100% CPU utilization:

https://i.gyazo.com/83046933d340b2fb3807456204085f45.png

Any reason for that? Throttling ?

[tp4tissue]

I have trouble getting the PC to run at 100% CPU utilization:

https://i.gyazo.com/83046933d340b2fb3807456204085f45.png

Any reason for that? Throttling ?

If that is the screenshot from prime running. Then there must be a power limiter somewhere preventing all core turbo.

As I said before , I'm not familiar with this chipset,   this may be by design (default) as you'd probably need a ~300 watt cooler to do run that chip full tilt. 

[Naweo]

I have trouble getting the PC to run at 100% CPU utilization:

https://i.gyazo.com/83046933d340b2fb3807456204085f45.png

Any reason for that? Throttling ?

If that is the screenshot from prime running. Then there must be a power limiter somewhere preventing all core turbo.

As I said before , I'm not familiar with this chipset,   this may be by design as you'd probably need a ~300 watt cooler to do run that chip full tilt. 

Say I won't find a solution to that, is it still fine to run the OC? I want to find/test for statbility, as mentioned.

[tp4tissue]

Say I won't find a solution to that, is it still fine to run the OC? I want to find/test for statbility, as mentioned.

The stability test will still be valid. "under" those conditions.

[Naweo]

Say I won't find a solution to that, is it still fine to run the OC? I want to find/test for statbility, as mentioned.

The stability test will still be valid. "under" those conditions.

Okay, I will try to find the right voltage and frequency with current settings then, despite the throttling

Thank you

[tp4tissue]

Okay, I will try to find the right voltage and frequency with current settings then, despite the throttling

Thank you

in HWinfo , the sensor mode, has a performance limit reason drop down, it should tell you what's happening, what variable is triggering the throttle.

[Naweo]

Okay, I will try to find the right voltage and frequency with current settings then, despite the throttling

Thank you

in HWinfo , the sensor mode, has a performance limit reason drop down, it should tell you what's happening, what variable is triggering the throttle.

You men HWmonitor or HWInfo.com?

[tp4tissue]

[Naweo]

(Attachment Link)

https://i.gyazo.com/b3c5bbd19cd4d0a0e45cfc713383aa47.png

It seems to be the case

[tp4tissue]

It seems to be the case

There's an option to disable it,  but You gotta make sure your cooler can handle all core turbo.   OR, you can set the turbo lower.

[Naweo]

It seems to be the case

There's an option to disable it,  but You gotta make sure your cooler can handle all core turbo.   OR, you can set the turbo lower.

Will try to find it, thank you.

And you mentioned that it's good for peformance/stability to disable?

[tp4tissue]

And you mentioned that it's good for peformance/stability to disable?

It allows you to test different multipliers separately to ensure that the voltage given to the cpu is correct/stable.

That setting itself may not be a stability determinant since your Final outcome may radically differ.

You may want to read more indepth guides on overclocking.  I can't explain the process in <bite size>,  It's a thick totality.

[Naweo]

No freeze for 10 days straight. I think that's a record so far, since I overclocked. Will keep posting.

[tp4tissue]

disable as many power saver settings as possible.   Like c-states.

[Naweo]

Update: monitors go black all the time now after reboot. 20 times a day

[tp4tissue]

Update: monitors go black all the time now after reboot. 20 times a day

/headscratch....  If you revert the motherboard tunings, is it better ?

[Naweo]

I had travelled and started messing with cables again.

I had a 5th, HDMI cable (plugged into PC but on the other end into nothing.)

Unplugged it. So far no crash for 10 hours.

Lets see if lucky.

Prior to travel I had a 12 day streak no crash.

[Naweo]

disable as many power saver settings as possible.   Like c-states.

For the CPU throttling: I found:

CPU TJ max: auto

Long duration power limit: Auto

Short duration power limit: auto

CPU current limit: auto

Anything to set manually to prevent throttling? Cuz its still doing it

[tp4tissue]

Look at the PL1 and PL2 power limit in HWinfo,  if it says something like 4 kilowatt, or something extremely high, then you don't need to change the duration power limits.

Those settings restrict/ cause throttling if auto means capping to rated sustained tdp. <this differs between boards> some auto means uncapped, some auto means capped

Maybe see if there's a setting called MCE,  multicore enhancement, with it enabled, the power limits <on auto> should be essentially off.

[Naweo]

https://i.gyazo.com/5efec18c63fdef661b8eb83dcdc03aba.png

https://gyazo.com/364b8ad0f485a5acdcff2dff91611dd7

This is what I found on my Hardwaremonitor

[tp4tissue]

that's hardware monitor,   you need to use hwINFO.

[Naweo]

Also: regarding black screen, it also happens randomly to any screen for 2-3 seconds, the bounces back, a few times a day during 10 hours sessions.

[Naweo]

that's hardware monitor,   you need to use hwINFO.

(Attachment Link)

Sorry, Thank you.

I found this:

https://i.gyazo.com/e424a1e1449ea9482c2dfc6380d7c434.png

Seems to be stuck at 200w. Low?

[tp4tissue]

Well it's not _low_ in the sense that it's by design, but that means you may need to enable mce somewhere.

Or it might require changing those limits.

Do testing with 2 sticks of ram, 4 sticks is iffy sometimes.

[Naweo]

so I think my GPU just died. I had a power outage and now it only loads one monitor that is powered by USB.

My GPU overclock asus software says "can't load vga driver."

I booted my PC once and it did five beeps prior to resulting in this error.

Is it possible the issue has been GPU all along?

Is there a way to check if my GPU is dead, if I don't have access to another PC right now?

[tp4tissue]

does the cpu have integrated ?

[yui]

so I think my GPU just died. I had a power outage and now it only loads one monitor that is powered by USB.

My GPU overclock asus software says "can't load vga driver."

I booted my PC once and it did five beeps prior to resulting in this error.

Is it possible the issue has been GPU all along?

Is there a way to check if my GPU is dead, if I don't have access to another PC right now?

monitor doing something funky can always be GPU, or anything feeding the GPU power or data, it is the problem with diagnosing those problems without spare parts.
you can search the meaning of the beep codes online but i find it strange that if the GPU is outright dead it managed to display anything. i would more expect the power outage to corrupt the windows install or drivers and that to cause them not to load, but then there are the beeps although most commonly 5 beeps mean CPU/motherboard failure (it is only CPU but i had more often dead motherboard with a good CPU report a bad CPU that actual bad CPU), then still without CPU it should not even bother to go further into POST and never should boot and try to load drivers, so unlikely that this code is right for your pc... maybe the bios somehow also got corrupted in the power failure? (very unlikely, very much more likely a completely unrelated error to CPU and you need to check online for your particular model, maybe a hdd secureboot failure if there was corruption to the OS for example, or short DST failure indicating a dying hdd/ssd (that could explain all sorts of completely random failures all over the place as windows could get more and more corrupted by bitrot)).
but yeah by coincidence your GPU could have been dying all along and so happen to outright fail on a power cut but then 30 series had capacitors and powerstages problems, and increasing the heat by overclocking would only make those problems worse (kinda why i never agreed with TP4 on overclocking being the solution but then not my system either)

[Leslieann]

I booted my PC once and it did five beeps prior to resulting in this error.

For Asus...
1 long and 4 short is a drive failure.
5 short beeps is a CPU failure.

While the drive failing plays into everything I said earlier about drivers the cpu failing plays into what you've been doing lately, which may have compounded your problem and now you're dealing with two problem not just one.