Hi all, thanks for trying to help, sorry for the long text
Original Specs:
ASUS Z-170P, Windows 10 Pro
Core i5-6500 @ 3.2 GHz, stock cooler
AMD RX 480 Founders Edition
2 x 8GB HyperX Fury DDR3-1866 DIMM 240
beQuiet! Pure Power L8 630W
So even when I built this thing couple years back I instantly started having problems with the GPU. Fans wouldn't turn on under load and it would overheat quick and lead to lags when gaming. I found a workaround using MSI Afterburners Auto settings, and while the chip would run quite hot (but not TOO hot) and the fan would run fast, it did the job for 3.5 years. I chalked it up to a faulty GPU because founders edition and just left it at that...
Performance deteriorated rapidly over the last few weeks and I started having random driver hang-ups and freezes. I decided to open up the rig and saw that it was significantly more dirty than the last time. Oops. I opened the RX480 and cleaned out big chunks of dust off the heatsink. I put it back together but after reassembling the rig the GPU started black screening (No Signal) pretty much at the start of any game or even opening the AMD Radeon driver software, and the fans would start spinning at what almost felt like 150%. Only way to stop the PC was to kill it with the power switch. I could hear that the PC was still working but there was no way to restart the video driver at this point
Figuring that I had just damaged the already faulty GPU even further and that it was just a matter of getting a new GPU. I managed to grab one of the few GPUs on the market, a XFX RX 580 GTS XXX Edition, and plugged it in. This time with the new GPU, I had the same error, with added weird artifacts before each crash. The PC doesn't completely crash now though and I get this message after driver crashes. The PC is working more or less but the video drivers still crash when trying to play any games or browsing through AMD Radeon driver software. It also seems to crash whenever the GPU or CPU load is ever so slightly challenging. The Performance tab in the AMD software, when it doesn't crash, shows me that the GPU load is at 100%, while Windows task manager shows only a few percent are being used..
Open Hardware Monitor though also shows 100% usage .
I can manage to get a game running in the main menu with extremely low fps for a minute or two before the game crashes, even if the video driver crashes during loading of the game. The performance tab shows me that the GPU load just spikes heavily during this before crashing
Weirdly enough, I seem to have the same issue as with the RX480. Without MSI Afterburner, the RX580 also seems to just keep heating and heating and the fans don't ever activate by themselves. With Afterburner, the card is cool and not too noisy.
I started suspecting a PSU issue so I bought a cheap but still sufficient PSU called LCPower LC6550 550W but that didn't solve the problem either. I've also replaced the stock CPU cooler with a beQuiet Pure Rock Slim for unrelated reasons, but I now think the CPU might be involved. Using the Event Viewer, I managed to see that a certain "amdkmpfd.sys" driver failed or was at the cause of the driver crashing whenever the crashes occur. Also, if I check the PCI slot in the Device Manager, it shows no issues. If I try to "Update Drivers" in the PCI slot properties though, it actually downloads and installs some drivers through windows, but after installing these drivers, the AMD software doesn't work and says it can't run with incompatible drivers installed.
Right now, I'm at a bit of a loss. I'm pretty confident I've ruled out software defects but I can't even tell. I have already tried:
I will know more tomorrow when I will be able to check the GPU(s) and RAM sticks in a friends computer, but until then, do you guys have any idea what the heck is going on or what I could try to troubleshoot further? It seems like a hardware fault at this point but, where?!
Edit: some more info:
If I try to run a game without AMD drivers installed, I get a BSOD with win32kbase.sys failing, this is the message in the log:
The computer has rebooted from a bugcheck. The bugcheck was: 0x00000050 (0xffff959ce8831fb8, 0x0000000000000000, 0xfffff8034281a74d, 0x0000000000000002). A dump was saved in: C:\WINDOWS\MEMORY.DMP. Report Id: 73177fb3-1122-4b19-b66f-ec2c9f682d3c.
This also happens at the second of the crash:
Driver Management concluded the process to install driver pci.inf_amd64_bd42a023f5b03760 for Device Instance ID ACPI\PNP0A08\0 with the following status: 0x0.
Original Specs:
ASUS Z-170P, Windows 10 Pro
Core i5-6500 @ 3.2 GHz, stock cooler
AMD RX 480 Founders Edition
2 x 8GB HyperX Fury DDR3-1866 DIMM 240
beQuiet! Pure Power L8 630W
So even when I built this thing couple years back I instantly started having problems with the GPU. Fans wouldn't turn on under load and it would overheat quick and lead to lags when gaming. I found a workaround using MSI Afterburners Auto settings, and while the chip would run quite hot (but not TOO hot) and the fan would run fast, it did the job for 3.5 years. I chalked it up to a faulty GPU because founders edition and just left it at that...
Performance deteriorated rapidly over the last few weeks and I started having random driver hang-ups and freezes. I decided to open up the rig and saw that it was significantly more dirty than the last time. Oops. I opened the RX480 and cleaned out big chunks of dust off the heatsink. I put it back together but after reassembling the rig the GPU started black screening (No Signal) pretty much at the start of any game or even opening the AMD Radeon driver software, and the fans would start spinning at what almost felt like 150%. Only way to stop the PC was to kill it with the power switch. I could hear that the PC was still working but there was no way to restart the video driver at this point
Figuring that I had just damaged the already faulty GPU even further and that it was just a matter of getting a new GPU. I managed to grab one of the few GPUs on the market, a XFX RX 580 GTS XXX Edition, and plugged it in. This time with the new GPU, I had the same error, with added weird artifacts before each crash. The PC doesn't completely crash now though and I get this message after driver crashes. The PC is working more or less but the video drivers still crash when trying to play any games or browsing through AMD Radeon driver software. It also seems to crash whenever the GPU or CPU load is ever so slightly challenging. The Performance tab in the AMD software, when it doesn't crash, shows me that the GPU load is at 100%, while Windows task manager shows only a few percent are being used..
Open Hardware Monitor though also shows 100% usage .
I can manage to get a game running in the main menu with extremely low fps for a minute or two before the game crashes, even if the video driver crashes during loading of the game. The performance tab shows me that the GPU load just spikes heavily during this before crashing
Weirdly enough, I seem to have the same issue as with the RX480. Without MSI Afterburner, the RX580 also seems to just keep heating and heating and the fans don't ever activate by themselves. With Afterburner, the card is cool and not too noisy.
I started suspecting a PSU issue so I bought a cheap but still sufficient PSU called LCPower LC6550 550W but that didn't solve the problem either. I've also replaced the stock CPU cooler with a beQuiet Pure Rock Slim for unrelated reasons, but I now think the CPU might be involved. Using the Event Viewer, I managed to see that a certain "amdkmpfd.sys" driver failed or was at the cause of the driver crashing whenever the crashes occur. Also, if I check the PCI slot in the Device Manager, it shows no issues. If I try to "Update Drivers" in the PCI slot properties though, it actually downloads and installs some drivers through windows, but after installing these drivers, the AMD software doesn't work and says it can't run with incompatible drivers installed.
Right now, I'm at a bit of a loss. I'm pretty confident I've ruled out software defects but I can't even tell. I have already tried:
- Reinstalling all drivers, deleting all using DDU, reinstalling windows and reinstalling drivers, etc.
- Thoroughly cleaning the motherboard with isopropyl alcohol and canned air the PCI slots
- sfc /scannow showed no issues
- Intel Diagnostic Tool for CPU show no issues
- switching RAM channels and alternating RAM sticks and Channels
- Reset BIOS / CLRTC and updated to newest version
- Switching to a different PCI slot
- Switching out the PSU
I will know more tomorrow when I will be able to check the GPU(s) and RAM sticks in a friends computer, but until then, do you guys have any idea what the heck is going on or what I could try to troubleshoot further? It seems like a hardware fault at this point but, where?!
Edit: some more info:
If I try to run a game without AMD drivers installed, I get a BSOD with win32kbase.sys failing, this is the message in the log:
The computer has rebooted from a bugcheck. The bugcheck was: 0x00000050 (0xffff959ce8831fb8, 0x0000000000000000, 0xfffff8034281a74d, 0x0000000000000002). A dump was saved in: C:\WINDOWS\MEMORY.DMP. Report Id: 73177fb3-1122-4b19-b66f-ec2c9f682d3c.
This also happens at the second of the crash:
Driver Management concluded the process to install driver pci.inf_amd64_bd42a023f5b03760 for Device Instance ID ACPI\PNP0A08\0 with the following status: 0x0.
Last edited: