Identifying reasons for PC black screen during operation?
Identifying reasons for PC black screen during operation?
Thank you for sharing the detailed information. Here’s a clearer version of your content while keeping the same structure and length:
Current specs:
Motherboard: ASRock X870E Taichi
CPU: Ryzen 9950X3D with AIO Cooler
GPU: ASUS 4090 TUF OG - Stock Cooler
PSU: EVGA Supernova G7 1000
RAM: G.Skill 64 GB DDR5 6000 (F5-6000J3040G32GX2-TZ5NR) with EXPO
I also experimented with F5-6400J3239G32GX2-TZ5RK (XMP 6400) RAM, but I wasn’t aware it wasn’t on QVL, so I decided to try one that was.
Recently, I upgraded my CPU, motherboard, and RAM for a 64 GB 9950x3d configuration. Overall, the experience has been positive, but recently I encountered Expedition 33 and began facing black screens and freezes. Since I regularly use Linux, I was familiar with NVIDIA problems, but now I noticed two new issues:
1. The game would freeze intermittently after a while in random areas. The fans slowed down, suggesting the DXVK chip might have failed. Video played normally in a web browser, likely due to software acceleration.
2. In the worst scenario, screens would go black unpredictably, monitors would shut off, GPU fans would be at maximum, system audio would still play, and I could connect via SSH. Linux logs indicated Xid 59 and "GPU has fallen off the bus." While Xid 79 usually points to thermal or power issues, temperatures stayed around 65-70°C.
To investigate if the problem was related to Linux drivers, I tested both open-source and proprietary drivers.
I also have a dual-boot Windows system, so I tried various driver versions—latest 576.15 and older 560.70. Across all versions, the black screen issue continued. Instead of freezing the game, I encountered a fatal error in UE5. To rule out other games, I ran Oblivion Remastered, which also crashed with a similar issue.
I’ve re-seated the GPU multiple times, replaced the 12VHPWR connector, and reconnected the PSU side. Turning off EXPO didn’t help. I even used my old 2070 graphics card, which gave somewhat similar results, though I didn’t test enough to confirm black screens. I suspect this isn’t a GPU problem, as I’ve owned the card for about two years without such issues. I purchased the PSU at the same time, so I doubt that’s the cause either.
From GPU-Z data, my GPU hotspot temperatures reach up to ~120°C, and PerfCap suggests it’s usually Thermals-related, though sometimes Vrel does. At this stage, I’m unsure what’s causing the problem, having tried many solutions. My old system used a 3700X AM4 socket, making it hard for me to test older components. Thank you for your help!
Your hot spot must stay below 120°C when the core operates at 70°C. This represents a 50° difference, significantly larger than the usual 15-20° range. It seems the card could have thermal connection problems and may require reworking. This isn't necessarily the root cause of your issue, but it suggests the card is currently under strain.
Welcome to the forums, newcomer!
Possible causes?
1) Driver problem
2) OS corruption
3) GPU overheating
4) Refresh rate mismatch with driver/software/monitor/GPU
I recently upgraded my CPU, motherboard, and RAM for a 64 gb 9950x3d build. Did you reinstall the OS after your platform change?
Motherboard: ASRock X870E Taichi
What is the BIOS version for your motherboard?
I also have a dual-boot Windows configuration, so I tried different driver versions—576.15 is the latest and 560.70 is the oldest.
It might help to use DDU to remove all GPU drivers (Intel, AMD, Nvidia) in Safe Mode, then reinstall manually with 566.36 via an elevated command, such as right-clicking the installer and selecting Run as Administrator. Check if the problem improves.
Regarding the OS, I plan to use a blank spare drive, install one copy of Windows offline, and then manually add all necessary drivers to test. Dual-boot systems often show surprises when issues arise.
Not originally, but I recently performed a complete Windows reinstall just in case.
Version 3.20 is the latest.
I’m going to try it and check if it resolves my problem.
Uncertain whether this would impact anything, but I have my Windows running on a separate 1TB SSD as my primary Linux drive. I believe they share the same EFI partitions, though I’m not sure if that could cause issues.
I’m also wondering if there might be a motherboard-related cause, since this also occurs with my older card?
Excluding the ASRock + 9800X3D problems, I’m curious if reseating the PCIe slot after multiple changes could have affected it.
Your hot spot must stay below 120°C when the core operates at 70°C. This represents a 50° difference, significantly larger than the usual 15-20° range. It seems the card could have thermal connection problems and may require reworking. This isn't necessarily the root cause of your issue, but it suggests the card is currently under strain.
Therm is the sole PerfCap reason that doesn't fit, and you should never see it. Supporting the repasting advice.
Alright, it's been a couple days, and it seems like repasting (I used PTM instead), has drastically reduced temperatures and no black screens or freezes since. I'll also probably looking into doing some undervolting, since I've heard the 4090 can go kinda low and still have relatively similar performance with less heat. Thanks all!