Question Games constantly crashing on custom build ?

[ElConquistadorDeAmor]

Spoiler: CLICK HERE TO SEE SPECS

CPU	Intel Core i9-13900K - Core i9 13th Gen Raptor Lake 24-Core (8P+16E) P-core Base Frequency: 3.0 GHz E-core Base Frequency: 2.2 GHz LGA 1700 125W Intel UHD Graphics 770 Desktop Processor - BX8071513900K	https://www.newegg.com/intel-core-i9-13900k-core-i9-13th-gen/p/N82E16819118412?Item=N82E16819118412
CPU cooler	be quiet! 250W TDP Dark Rock Pro 4 CPU Cooler with Silent Wings - PWM Fan - 135 mm LGA 1700 Compatible	https://www.newegg.com/be-quiet-dark-rock-pro-4-bk022/p/13C-001F-00027?Item=9SIA68V6YA3005
Motherboard	GIGABYTE Z790 AORUS ELITE AX LGA 1700 Intel Z790 ATX Motherboard with DDR5, Quad M.2, PCIe 5.0, USB 3.2 Gen2X2 Type-C, Intel WiFi 6E, 2.5GbE LAN, Q-Flash Plus, PCIe EZ-Latch	https://www.newegg.com/p/N82E16813145417?Item=N82E16813145417
RAM	G.SKILL Trident Z5 RGB Series 32GB (2 x 16GB) 288-Pin PC RAM DDR5 6400 Desktop Memory Model F5-6400J3239G16GX2-TZ5RK	https://www.newegg.com/g-skill-32gb-288-pin-ddr5-sdram/p/N82E16820374358?Item=N82E16820374358
GPU	GIGABYTE Gaming GeForce RTX 4090 24GB GDDR6X PCI Express 4.0 x16 ATX Video Card GV-N4090GAMING OC-24GD	https://www.newegg.com/gigabyte-geforce-rtx-4090-gv-n4090gaming-oc-24gd/p/N82E16814932550?Item=N82E16814932550
Case	Thermaltake Core X71 Tempered Glass Edition Black ATX Gaming Full Tower Tt LCS Certified Gaming Computer Case CA-1F8-00M1WN-02	https://www.newegg.com/black-thermaltake-core-x71-tg-edition-atx-full-tower/p/N82E16811133331?Item=N82E16811133331
PSU	CORSAIR RM1000e Fully Modular Low-Noise ATX Power Supply - ATX 3.0 & PCIe 5.0 Compliant - 105°C-Rated Capacitors - 80 PLUS Gold Efficiency - Modern Standby Support	https://www.newegg.com/corsair-rm1000e-cp-9020264-na-1000-w/p/N82E16817139305?Item=N82E16817139305
Hard drive #1 (M2 drive)	SAMSUNG 980 PRO M.2 2280 2TB PCIe Gen 4.0 x4, NVMe 1.3c Samsung V-NAND Internal Solid State Drive (SSD) MZ-V8P2T0B/AM	https://www.newegg.com/samsung-2tb-980-pro/p/N82E16820147796
Hard drive #2 (SATA drive)	SAMSUNG 870 EVO Series 2.5" 4TB SATA III V-NAND Internal Solid State Drive (SSD) MZ-77E4T0B/AM	https://www.newegg.com/samsung-4tb-870-evo-series/p/N82E16820147795?Item=N82E16820147795

I checked my BIOS and this is the relevant info:

• Model name: Z790 Aorus Elite AX
• BIOS verison: FC
• BIOS date: 03/07/2023
• BIOS ID: 8ARPT050

---------------
Hey guys, I'm having a problem with a PC I built in September start to have increasingly severe problems with games crashing, usually on loading levels or on launch. This was not a problem until the last couple months, so I don't think it's because I made mistakes in assembling the hardware. With my specs, I should not be having any problems running games. There is a lot going on here, so I'll try to digest this into easily readable bullet points.

Generally speaking, usually games crash to desktop, but sometimes the entire system reboots. I don't have any problems outside of games, except some extremely minor artifacting in my browser on occasion that only lasts a second or two, which I would say is relatively recent. The games all run really smoothly with high framerate until they crash. Here are the problems I'm currently having with the primary games I'm playing.

• Age of Wonders 4: can usually play around 30–60 mins before a crash, but I'm noticing that crashes become more frequent the more I reboot the game. AoW4 is not a particularly demanding game, so this is very frustrating. I initially though this was just the game or was mod-related, but now I suspect it's related to my other crash problems.
• Ready or Not: sometimes I get a crash on launch with an error that sufficient video memory cannot be allocated, sometimes I get a crash on loading a level with no error. Sometimes I can play for hours with no problem. Performance monitors don't show anywhere near my VRAM capacity being reached.
• Cyberpunk 2077 & Metro Exodus: reliably crash every time I boot the game within a few seconds. No error messages. I specifically checked these games when I started noticing issues since I used to be able to play them with zero issue.
• Helldivers 2: can boot into the game and run around the ship, but as soon as I launch a level or boot the tutorial, the game crashes every single time with no obvious error messages. Game is literally unplayable.
• Lethal Company: usually is fine, but loading a specific level (Titan) seems like a roll of the die whether it will crash my whole system. Maybe like a 20% chance that Titan just crashes to desktop or reboots my whole damn system.
• Darktide: I seem to have no problems with this game at all despite it being a relatively demanding game.
• Hitman: Usually I can load to the main menu, but as soon as I boot a mission, it crashes. Sometimes it crashes immediately, but usually I can at least get to the main menu.
• Lower demanding games, usually older games don't seem to have much problem

Here are the things I've tried already:

• Uninstalled my GPU drivers with Display Driver Uninstaller (DDU) and reinstalled the latest versions
• Downloaded both GPU-Z and CPU-Z to monitor temps and performance in real time. I see nothing of note.
• Watched the Windows Task Manager's performance monitors. Again, nothing of note.
• Run a stress test at max load for 40 minutes. No problems at all.
• Run a standard memory test using Windows Memory Diagnostic. No errors found.
• Looked at Event Viewer for system messages when applications crash. Maybe I don't know what to look for, but I haven't found anything useful.
• Checked for whether the culprit games are running on a specific SSD, but this seems to not be a factor considering both my M.2 drive and my other internal SSD have games running on them that crash.
• Ran a stress test on Hitman specifically using my onboard graphics card, and the same crash happened. I really hope this at least means my GPU is not the culprit.

I would prefer not to do invasive stuff like this until I have to, so here are things I have not yet tried:

• Done a clean install of Windows 11
• Anything BIOS-related
• Opened up the case for examination
• Swapped out any hardware components

Can anyone help? I am going absolutely crazy with this. I have no idea what the problem could be. I suspect it's a hardware failure somewhere, likely the memory or PSU given when the crashes happen, but I truly am at a loss here.

EDIT: Please see one of my posts down below where I got my games to work by underclocking my CPU. I still think this is an issue.

 

[Nine Layer Nige]

Please provide a full list of your components make, model.
BIOS version
PSU is a probably cause of instability

 

[ElConquistadorDeAmor]

Please provide a full list of your components make, model.
BIOS version
PSU is a probably cause of instability

Good point; I've updated the OP with full specs, and I included BIOS information. Click the spoiler at the top to see it.

 

[Nine Layer Nige]

Are you running XMP ?
Are you manually overclocking ?

 

[ElConquistadorDeAmor]

Are you running XMP ?
Are you manually overclocking ?

I’m neither running XMP nor have I done any manual overclocking. Truth be told, I really haven’t fiddled with anything performance-related like that on this computer.

 

[Nine Layer Nige]

So what speed is your ram running at then ?

 

[ElConquistadorDeAmor]

So what speed is your ram running at then ?

I will check this tomorrow for you. I had to Google what XMP is, so if that’s not something one needs to do manually, I’ll also try and double check whether my system is actually using it or not.

 

[Nine Layer Nige]

Generally, if you have built the system yourself, you would need to switch XMP on in the BIOS and it will automatically find the fastest most stable frequency to run your ram at. According to the Gigabyte website, your ram natively runs at 4800mhz but is capable under XMP of running upto 6400.

If it was my rig, I would update the BIOS to the lastest version and then switch on XMP and see what happens. However, every BIOS update has a risk, so consider that.
As I mentioned, even if a PSU is new, it can create stability issues, so if you can borrow a PSU to test, that might help diagnose the cause.

 

[NedSmelly]

I disagree with the XMP suggestion. IMO for troubleshooting it makes more sense to leave the RAM at standard (JEDEC) settings. Turning on XMP just adds another troubleshooting variable in terms of stability.

OP: There have been some reports of RTX4090s drawing spikes in overcurrent which can overwhelm the PSU and trigger the OPP safety threshold. That might be a path for further investigation. My gut feeling is a GPU issue - especially when you’ve checked all those other things in your list. That “minor artifacting in my browser” is somewhat ominous.

From MSI: “…a 450W RTX 4090 can have a transient spike of up to 1350W”

EDIT: I’d also run Memtest86 overnight. It’s more thorough than Windows Memory Diagnostic.

 

[Nine Layer Nige]

I agree. Maybe my XMP comment may not help to start with.

 

[PEnns]

Looks like a solid system.

But: Between the i9-13900K and the GeForce RTX 4090, the 1000 W PSU may not be enough, unfortunately.

 

[NedSmelly]

Looks like a solid system.

But: Between the i9-13900K and the GeForce RTX 4090, the 1000 W PSU may not be enough, unfortunately.

Good point. If the OP is comfortable with BIOS changes they could temporarily limit the CPU’s PL2 to 200W or less and PL1 to 125W, to contain total power consumption as a stability test. Maybe they have one of those motherboards that default to “unlimited power” to the CPU.

 

[ElConquistadorDeAmor]

Hmm, I was really hoping the GPU wasn't the problem. But given the relatively new artifacting I've been noticing, I have my suspicions. I really suspect this must be a hardware issue somewhere between the PSU, GPU, and memory. I unfortunately don't have a spare PSU laying around anywhere near powerful enough for this system to swap out for a test. I wouldn't have thought the PSU's total 1000W power would be the problem given it hasn't posed any problem until roughly around last month, but who knows?

For the record, the artifacts that I have noticed don't look quite as severe as the ones I see on Google image search. It's more just some gray squares in the upper left corner now and then. But that's probably irrelevant and it is in fact artifacting.

I disagree with the XMP suggestion. IMO for troubleshooting it makes more sense to leave the RAM at standard (JEDEC) settings. Turning on XMP just adds another troubleshooting variable in terms of stability.

OP: There have been some reports of RTX4090s drawing spikes in overcurrent which can overwhelm the PSU and trigger the OPP safety threshold. That might be a path for further investigation. My gut feeling is a GPU issue - especially when you’ve checked all those other things in your list. That “minor artifacting in my browser” is somewhat ominous.

From MSI: “…a 450W RTX 4090 can have a transient spike of up to 1350W”

EDIT: I’d also run Memtest86 overnight. It’s more thorough than Windows Memory Diagnostic.

The GPU doing power spikes makes intuitive sense to me based on the symptoms I'm seeing, since these crashes usually occur on loading levels. At any rate, I'll try your suggestion of running Memtest86 and report the results.

Good point. If the OP is comfortable with BIOS changes they could temporarily limit the CPU’s PL2 to 200W or less and PL1 to 125W, to contain total power consumption as a stability test. Maybe they have one of those motherboards that default to “unlimited power” to the CPU.

I am willing to try that at some point if I have to, but I would like to avoid BIOS changes if I can help it. I will keep this in mind, however.

I really appreciate all the responses so far, guys! Thank you very much.

 

[ElConquistadorDeAmor]

Okay, I've just made somewhat of a breakthrough, though I'm not sure I'm entirely happy about how I "fixed" it. I actually wound up enabling XMP 1 and tried launching my games again. Metro Exodus worked, but Cyberpunk still didn't. I checked the crash log for Cyberpunk and it said EXCEPTION_ACCESS_VIOLATION (0xC0000005). I looked this up and it seems like it's maybe a DirectX issue, which I suppose makes sense because I've had more difficulty playing Ready or Not with DirectX 12 than DirectX 11.

I then went back into my motherboard setup and underclocked my CPU's base clock speed from Auto (100.00 Mhz) to 86.00 Mhz. Suddenly everything seems to be working. Metro Exodus and Cyberpunk both launched, and I even got Helldivers 2 to successfully launch a mission. That's just been my experience in the 15 minutes of testing that I've done since making this change.

Obviously I'm not entirely happy, because underclocking my CPU is going to slow down my performance a bit. That said, I can only assume this points to some underlying issue somewhere. Can anyone help diagnose what must be going on under the hood here? I should not have to neuter my CPU to make my games start working again.

I did try to find the power consumption settings for the CPU like NedSmelly suggested, but I didn't see them anywhere. Maybe that's in the actual BIOS setup instead of the setup menu I bring up with F12 when the PC is booting.

 

[NedSmelly]

I did try to find the power consumption settings for the CPU like NedSmelly suggested, but I didn't see them anywhere. Maybe that's in the actual BIOS setup instead of the setup menu I bring up with F12 when the PC is booting.

In BIOS go to ‘Advanced Mode’, Tweaker, Advanced CPU, Turbo power limits, Power limit TDP / Power limit time

Return base clock to standard before experimenting with turbo and base power limits. Note that you need to select a power limit time (Tau) with K processor otherwise mobo will stay in turbo indefinitely. Used to be 56 seconds in older K CPUs

PS I used to get exception access violations with CP2077 until I discovered I had a faulty stick of RAM using Memtest86

 

[ElConquistadorDeAmor]

I was a bit lazy and just dealt with my Band-Aid solution of keeping the CPU clock speed down since it "fixed" my issue. I'm coming back to this thread now because my friend showed me this video that suggests Intel 13th/14th Gen Raptor Lake CPUs are frying themselves after 1–3 months of use and will make games crash when compiling shaders. Interestingly, the error "Out of video memory trying to allocate a rendering resource" is something I used to get. Do you guys think this could just be a hardware issue with my CPU then?

My same friend also sent me this Reddit thread suggesting a lot of motherboards reroute too much power to this particular CPU. I will have to try this at some point and let you all know if it fixes it.