First of all, sorry about my english but i'm from Spain, so i'will try to explain myself the best i'can.
Recently I'decided to build a new gaming PC, so i'bought some new components and i'have installed another ones from my old build. The result is the following one:
Processor: Amd ryzen 7 3700x with stock cooler and stock settings (new)
Mobo: Gigabyte aorus x570 ultra (new)
Ram: G skill trident Z neo 16gb 2x8@3600mhz cl16 (model number f4-3600c16d-16gtznc) (new)
PSU: Antec high current gamer 850w 80 plus gold (new)
GPU: Powercolor rx 5700xt red devil (new)
Hard disks: SSD: samsung 950 pro nmve 240gb (from the old rig) wdc blue hdd 1tb (from the old rig too)
Case and fans: Lian li pc011 dynamic razer edition with 6 cougar argb fans (both new)
The problems started when i'first install my new components. First of all, i'installed new gpu (powercolor rx 5700xt , the two fan model, not the red devil one) with a new mobo (asus prime x570-p) and new processor (ryzen 7 3700x), with my old 750w psu from mars gaming and my old sdd and hdd. Problems start here, my pc was stuttering gaming and in windows, but after a while, the computers started working fine until you reboot it. I'thought that the problem was the gpu, so i'decided to rma it and pick the red devil edition model, the one that i'have now installed. Meanwhile, i'installed my old nvidia 970 with my new ryzen, motherboard and ram, and it worked perfectly without any issue, no random restarts, no stuttering...
When my new gpu arrived, the stuttering issue was totally solved, but the actuall issue appears. After some time gaming , the computer randomly restarts without dropping any bsod, it simply turns off and restarts the system itself. The problem don't have any type of time pattern , it can happens 2 or 3 times in a row and then works fine a entire day, so it's completely random. I'looked in windows event viewer and it drops kernel-power 41 error.
When this started to happen, i'thought that the problem was a faulty psu, so i'bought a new one (antec hcg 850w) but the problem was still there. Looking that it doesn't fix the issue, i'rma the cpu thinking it'was faulty and i'also buyed a better mobo (gigabyte aorus x570 ultra, the actual mobo installed on my setup), but the problem still persists.
So, with all that components replaced, i'started to search in the internet looking to fix the issue, and i'had tryed the following fixes, but saddly no one worked:
-Changed windows power plan to amd ryzen high performance (changing hdd turn off option to never and pci-e option too )
-Dissabled all hardware acceleration in all background programs like discord, battle.net, steam...
-Tryed pcie 3.0
-Tryed with precission boost overdrive on and off
-DDU all amd drivers and installed both 20.2.2 and 20.4.2 versions, performing a clean install, and using windows safe mode, and dissabling automatic driver updates from windows
-Dissabled radeon anti lag, radeon image sharpening, radeon enhanced sync and radeon freesync
-Downclocked ram to 3200 and 3000 mhz and setting to 1.35v
-Tryed xmp on and off
-Tryed memtest and hdd/sdd tests, both fine
-Tryed undervolting gpu
-All temperatures are ok, idle cpu 34º and gpu 53º and in stress test cpu 65º and gpu around 70-75º
-Tryed prime95 with furmark for hours , didn't crashed
-Dissabled fast boot from windows and bios
-Updated bios to gigabyte's last bios (aegesa f12e)
-Updated ssd firmware using samsung magician
-Updated to lastest chipset drivers and installed all amd sata drivers
-Changed thermal compound to cpu, despite it was freshly installed (thermal gryzzly)
On the event viewer, it always shows the kernel power 41 (63) error , but some lines before the kernel-power 41 event, it always show the following event: "Ntfs (Microsoft-Windows-Ntfs) id 98" (Volume C: (\Device\HarddiskVolume4) is correct. No action required.)
This event looked souspicious to me, so i'had check both my ssd and hdd, but they seem to be OK according to crystaldisk and chkdsk.
It also display some events after the kernel-power 41 event, that seem a bit souspicious and look related to the cpu. It drops 16 events named "Kernel-processor-Power (Microsoft-Windows-Kernel-Processor-Power) id 55 (47)
I'can also afirm that the reboots only happen while gaming, because i'never experienced the issue browsing the internet or at desktop. When stress testing the computer with prime95 and furmark , it works perfectly stable too.
I'beg you help guys, because i'cant find more information on the internet about the issue and i'm getting a bit frustrated.
I'will post any additional information required to solve the problem.
Tones of thanks
Solved! Go to Solution.
I bought an used gtx1070 just to test it. And as I suspected, no crashes, no freezes, no stuttering. I'm sending my 5700xt back for repairs/replacement. But I really hope the seller will be on top of their game and just take it back and return my money so I can buy rtx 2070/2080 super instead.
The Kernel power 41 error is just a notification that your system shut down incorrectly.
You seem to have checked all the common issues (power, heat, memory)
Type CMD into search
right click on command prompt app and run as admin.
enter sfc /scannow command.
This should scan for and fix any windows errors
enter chkdsk /f /r command
This will restart your computer and scan your drive for errors.
Thanks for your reply. As you said, i'executed scannow and chkdsk.
If a new crash happens i'will update.
Thank you a lot
Yesterday i'was able to play all day with no issues following the steps you provided and with xmp off. Today i'decided to put on xmp and give it a try, and just before entering world of warcraft, when the game was starting , it crashed again, but this time i'could hear the game playing at the background with a black screen before it restarted. It droped a volmgr 161 error in the event viewer, never saw this error in previous crashes.
I'look for memory compatibility list in g.skill website and my sticks are compatible to my mobo.
I'll try to manually oc the memory using dram calculator for ryzen anywais.
Lot of thanks
It happened again with xmp off. This time it didn't drop any volmgr 161 event. I'take a look to the device manager and the system itselft installed again realtek and amd audio drivers. I'had those unistalled yesterday because some people said that they could be the source of the issue. I'm trying again without them on, and will update the post again.
It sounds like it could be a stability issue perhaps due to unstable settings. I would run everything at stock, BIOS optimized default settings, even with your memory running at 2133Mhz auto timings until you can pinpoint the cause of your issue. Just from reading what you have written, it could be related to system memory settings.
It crashed again while gaming with bios stock settings, with memory running at 2133. I'run memtest 86 again and no errors where found. I'will search if i'can get more info in the internet.
Lot of thanks!!
It has nothing to do with memory , psu but with the drivers / radeon software that are basically crashing the entire system. I am having the same issue since last year and no driver updates managed to fix this.
Ryzen 5 3600X
Gigabyte X570 Elite
32 DDR4 (Corsair)
1 TB SSD
Sapphire 5700XT Nitro+
Corsair AX760 PSU
While in some games I have absolutely no issue playing for hours in others it just completely crashes the system or crashing to desktop. It's not the processor temp (65 at full load ), it's not the ram (no issues in 2-3 hour stress tests) , it's not the PSU (absolutely no issues in stress tests) . If it was the psu or mem , other games wouldn't have worked.
There is one game I can reproduce this constantly like GTA5. If I run the game in window borderless mode ... after some time it will crash to desktop. If I ran the game in full screen mode ...it will crash faster than in full window borderless and will also completely crash the system with an auto restart. Everytime I see that is it the radeon software restarting itself and system logs pointing it.
It is really sad that AMD is still uselesly going down the "PSU , mem , CPU overheating" troubleshooting path rather than admiting that there is something wrong with their software and it's a blow they still haven't got at the bottom of it since almost a year now.
Sad and will most likely be my last AMD video card. I should have paid the nvidia price and save me months of useless troubleshooting. Learned my lesson.