New gaming PC seriously underperforming and occasionally crashing

[313411]

I figured I might aswel try asking for help here as I was not getting many responses over on Reddit.

I let my PC get build and tested by a company where I picked the parts myself so I could finally have a good gaming rig. Now when playing games, I have noticed that it doesn't perform as well as I hoped / it should. (Fallout 4: 60FPS or lower, Rainbow Six Siege: Depending on startup either 20FPS or 90FPS, GTA 5: sometimes it starts off at 60+ but dips to 20 / 30 after 15 minutes. That or it starts at about 25FPS) And after a random amount of time while playing a game ranging from 30 minutes to an hour and a half, my game / computer completely freezes up while audio still works for about 5 to 10 seconds, then becomes choppy and then stops. CTRL + ALT + DELETE doesn't work at that point, neither does ALT + F4. Then after a minute or two the game completely shuts off as if nothing happened

Specs:

• 
  i7 8086k (not overclocked)
  16 GB DDR4-3200 RAM
  AORUS Gaming 7 motherboard
  AORUS GeForce GTX 1080 Ti Xtreme Edition 11G
  750 Watt PSU
  Windows 10 Home edition

The highest temperatures I've seen my CPU get was 60°C and for my GPU 61°C.

I have updated and later reinstalled my Nvidia drivers, checked Windows for updates and checked for any BIOS updates but everything is up to date and the issues still persists.

I'm completely stumped as to why my PC is under performing and crashing as much as it is doing and I hope any of you have an answer.

 

[Darkbreeze]

750 Watt PSU

This, is not enough information to make any kind of rational decision or determine whether the unit might be the source of the problem or not. MODEL number is required. Look ON the PSU, or the box it came in, and find the exact model number and brand of the unit. Without knowing that, we are just guessing at any kind of an answer.

9 times out of 10, on an entirely new system build issue here we see an older, used, cheap ass power supply in use and it is usually the problem. Not always to be sure, but a LOT of the time. If that unit is very old, or is a bargain basement type unit, it will likely be supplying maybe half it's labeled power and will likely be THE issue, or at least a contributor to it.

 

[313411]

My apologies, its a Be quiet! Straight power 11. Model E11-750W

 

[Calvin7]

Core i7-8086K
Supported memory: DDR4-2666

DDR4-3200 is for overclocking, but your not overclocking. If left on [Auto] memory settings, the memory may be running at a slow (SPD) Serial Presence Detect speed of 2133MHz.

 

[313411]

So am I basically screwed and do I have to buy different memory, or is there a way to salvage this?

 

[Calvin7]

What is the memory currently running at? Do you have the part number or link to it?

You can set memory parameters in BIOS manually or by enabling XMP (eXtreme Memory Profiles).

 

[313411]

Sorry if I am going to sound dumb, but where can I find that?

 

[Darkbreeze]

The memory configuration is AN issue, and I agree that it would be wise to enable XMP or manually configure the memory to take advantage of the speed it is assuredly not taking advantage of now, however, it is very unlikely that it is THE problem you are seeing.

Memory running at 2133mhz instead of 3200mhz is not going to cause the system to be freezing up and honestly, for almost everybody, they'd never actually be able to feel the difference in any meaningful way between the two speeds as 2133mhz, with a low latency, is already ridiculously fast. Bandwidth is not the bottleneck when it comes to memory performance on DDR4. Especially if you are running two or four modules and they are operating in dual channel, which effectively doubles the bandwidth.

BeQuiet has both decent, and horrible, power supply lines. Some of them are very good, others I wouldn't use any more than I would something from Ultra or Logisys, which are both dumpster fire quality. Obviously, BeQuiet isn't quite that bad, even on their inferior models, but without ANY reputable professional reviews of that unit it's hard to say one way or the other.

Probably not the issue, but I'd keep it in the back of my mind just in case.

What company built this for you? CyberpowerPC? iBuypowerPC? Somebody else?

Did you do the operating system installation, or did they? Have you checked to verify in Windows that all 16GB are being recognized and utilized?

What kind of drives are installed and what is the OS installed on?

Have you done a CLEAN driver installation, using the DDU, as outlined here:

http://www.tomshardware.com/faq/id-2767677/perform-clean-install-video-card-drivers.html

If you can't pin down anything else, I'd recommend, highly, doing a clean install of the OS, following the instructions at the following link exactly as outlined.

http://www.tomshardware.com/faq/id-3567655/clean-installation-windows.html

And THEN do a clean install of the GPU card drivers as outlined at the link above.

 

[Darkbreeze]

Sorry if I am going to sound dumb, but where can I find that?

In the bios, in the memory configuration. Sometimes in the CPU or tweakers tab. Your motherboard user manual should tell you where.

 

[313411]

I looked at it before, but I'll mark it down and let you know what it says in the BIOS later today once I get home from work. In the weekend once I have time I'll perform a clean install of Windows 10. The current version has been put on there by the company that built it for me which was Alternate, a company in Holland.

I have however checked if my computer recognized the 16GB of RAM and it did, luckily.

Thank you guys for your help so far!

 

[LordVile]

Have you installed all the drivers and checked you’re not running super sampling and running at 12K or something stupid.

 

[313411]

Okay so, in my BIOS, as far as I can remember most settings for my memory was set on auto, and the speed of which my memory was running (I assume) was set on 3200 but it was grayed out. I have done a clean uninstall / install with DDU in safemode, but today while I was playing House Flipper I started out with ~144 FPS, then suddenly tanked to somewhere between 20 and 35 and I completely froze again for about 6 minutes.

 

[Darkbreeze]

Did you happen to monitor or check on what your GPU thermal diode and CPU core temps were at the time that started dropping?

I'd install HWinfo and do that. There might be a relationship between thermals and the drop. Could be some kind of throttling.

Also, make damn sure that Windows system restore and degragmenter are both not set to automatically do ANYTHING, EVER. Disable them both. System restore never works right anyhow. Much better off with something like Acronis true image or another image backup utility that you can manually create images and not have it suddenly start working while you're trying to do something else. Same thing for disk defragmenter. Turn that off and periodically defrag or trim manually.

 

[313411]

During a drop my GPU temp was at 60°C, so it didn't really change.
According to HWinfo my memory is clocked at 1600 MHz however.

Edit: And just now while playing Sea of Thieves at 30 FPS, my GPU was at 53°C when it crashed.

 

[Darkbreeze]

Double data rate (DDR) at 1600mhz would be 3200mhz, which is right.

Please do the following. Download and run Furmark, the Heaven benchmark or another GPU utility that puts full stress on the GPU card. Run it in full screen mode BUT with HWinfo open on top of it. Take screenshots of the system voltage sensors in HWinfo, 12v, 5v, 3v, and post screenshots of those here. You will need to host them at imgur or tinypic, or another image hosting site that allows you to link to the actual image file ending in an image format like .jpg, .png, etc. and then use the following method to make it show up here.

You don't have to use the edit tools to do it, just use the same code you see below but insert the URL of your image in place with the two image tags at each end of the url.

 

[313411]

This one was when I started, for some reason the screen of FurMark didn't update on the screenshots. HWinfo was running for a couple of minutes before I started the benchmark. Screenshots are going in chronological order.
The imgur album (I don't know if I can post links): /a/2Eja40K

 

[Darkbreeze]

Hard to see, usually we'll edit out everything but the actual sensor part of the screen, but it looks like your voltages are fine going off the sensors. Not always 100% accurate, but usually at least a good indicator whether something is "off" or not in that regard. I think it's fine. I don't think the PSU is the issue here.

If you have the latest bios version installed, a clean install of windows, a clean install of the graphics card drivers and the most recent motherboard chipset drivers installed from the product page or the Intel website, then I'd say you've either got a faulty GPU card or motherboard. Obviously, there could be other things going on, but I'm beginning to suspect one of them.

What kind of drives are you using?

 

[313411]

Okay so I thought that a clean install of Windows + drivers fixed it, but I am right back at abysmal framerates.

What exactly do you mean by what kind of drives I am using?

 

[Darkbreeze]

I mean, hard drives, SSDs, M.2, etc.?

Do me a favor, and yourself I suppose, and download Prime95 version 26.6. Install it, run the Small FFT option. Take screenshots of the HWinfo sensors. Again, one screenshot of the HWinfo window, then scroll down, capture the next set of sensors, scroll down, capture the rest. However many screenshots is necessary to capture all of the various sensors and data.

But, this time, edit the image and only post the HWinfo screen, so it can be read easily. I have a hard time seeing the values in those screenshots you took.

Should look like this (Except, you'll have two or three screenshots, and will have different sensors showing. Mine is customized to show only the sensors I want to see):

 

[Darkbreeze]

We're talking about freezing up and unresponsiveness, not stuttering. I don't believe that any steam overlays or windowed gaming should cause any of those symptoms, at least, not unless something else is seriously wrong.

 

[smorizio]

as the pc was shipped to you. start with the power off make sure the ram and gpu and cpu are locked in. if the cpu not sitting on the pin issue may happen.
https://www.gigabyte.com/Motherboard/Z370-AORUS-Gaming-7-rev-10#support-dl-bios
if you have the 1.0 mb make sure you have the last bios f7 installed.
also from intel make sure you have the newest intel chipset drivers installed.
with the mb guild make sure the ram in the two main ram slots.
make a bootable memtest usb stick run it over night see if there any issues. with the mb guild and cpu-z check that your ram is on the mb qal tested list. with cpu-z under memory tab see what the voltage is needed to run at non stock speeds you may have to bump up the dram voltage and or change the timmings. to see if it ram issue turn off xmp set the ram t0 2100/2400 stock speed see if any of your issues stop. also download intel cpu test make sure the cpu memory controller is fine. use msi afterburner try under clocking the gpu speed and it ram speed. see if there an issue with the gpu.
last would be to try a test power supply see if your stock power supply not holding within atx spec.

 

[313411]

I currently have a Samsung 860 EVO 500 GB SSD for my OS and WoW (2,5 inch, I forgot to check how big it was... Stupid of me, but it fits), and a Seagate SkyHawk 4 TB HDD for the rest of my games.

Here are the screenshots I took while running Prime95 version 26.6, Small FFT.

@bimaloy30 this also happens in Sea of Thieves which is from the Microsoft store, and GTA V which I got from the Rockstar Social Club.

@smorizio I will give that a try tomorrow once I have the time for that. Not completely sure how to update my BIOS version though, have to look that up.

 

[Darkbreeze]

One of your fans, whichever is connected to the system_1 header, is shutting off. Probably don't want that happening. You don't want any case or cooling fan to ever be running at less than maybe 500rpm. Not that that is your problem, but you might want to take a look at that.

Your GPU temperature is 75 degrees???? With NO GPU stress running at all? Probably want to look at that too. Prime95 does not in any way stress the GPU card, so unless you were running something else simultaneously that was imposing a major load on the GPU card, something is definitely wrong there too. Make sure the GPU card fans are ALL working.

Your GPU thermal diode shouldn't show those kinds of temps unless it's under a serious load. Based on the sensor values showing 0% load, there is a thermal problem with the graphics card.

I would try a clean install of the GPU card drivers if there are newer drivers than when you did it before. If not, I'd contact them about an RMA. Something is definitely wrong with the GPU card thermals.

 

[313411]

I was playing a game an hour or so before this, so maybe because of that? I have noticed that my GPU fans barely even turn on unless I go into some settings to either activate them, for example in the AORUS Engine, they always automatically start at the Semi Passive option instead of at Active Fan, which means they never really turn on as far as I know since I believe that it automatically reverts to Semi Passive when I close the program. I turned on the fans with MSI afterburner for a moment like half an hour ago and its sitting at 45 degrees now.

I was kind of shocked to see those temperatures too however. Its the first time I've seen them this high! And for so long aswell.

 

[Darkbreeze]

Try setting a custom fan profile. Again, this might not be the problem, but I think it is. If it was at 75 degrees C at full load, I wouldn't be concerned. But 75 degrees with ZERO load, that is a major problem in my book, and at some point is probably leading to thermal throttling of the GPU.

Download and run the Heaven benchmark and simultaneously take a screenshot of the HWinfo GPU sensors, so we can see what it's actually doing under a load. Probably a good idea to open HWinfo first, so that if the GPU starts throttling right away once you open the benchmark you'll see it happen.