Is my Ryzen CPU defective ?

[TechnoStyle]

I have a problem with games. My AMD built PC (RX480 GPU & Ryzen 5 1600) is crashing at GTA IV, Metro Last Light, Metro 2033, Crysis1, Crysis Warhead, Dirt3: Complete Edition...

The interesting thing is, when PC freezed/ crashed, RGB LEDs of my graphics card is lighting red to white colours. It shows an error i think.
But also Ez debug led (CPU) of motherboard is indicating. (According to manual, it shows CPU faulty or no CPU found.)

Here is my topic about my GPU : https://community.amd.com/thread/216641
I tried different methods but couldn't solve the problem unfortunately.
So i don't know if my CPU or GPU or both are defective...
NOTE: CPU didn't overclocked. Stock speeds...
Today i tested my CPU with these methods:

AIDA64 stress test, 30 minutes --> NO CRASH.
I run Cinebench R15 for 10 times with this method: When rendering ends i run it again immediately before the CPU gets cold. --> NO CRASH.
I converted 885 mp3 files in 15 minutes to 320 kbps and again NO PROBLEM.
I run OCCT stress test with default settings for 1 hour --> NO ERROR.
I run HWinfo at last 30 minutes of OCCT test. --> 0 WHEA error.
I rendered 1.5 Gigabytes video. I set highest bitrate etc. and the coding is H265 HEVC. It took 3 hours and 15 minutes. File size of the result video is 6 gigabytes. --> NO PROBLEM/ NO CRASHING. Utilization is nearly %100 for all 12 threads.
I can surf on the web, watch Youtube videos etc. and didn't see BSOD.

So, my stock speed CPU is stable/ non-defective ?

 

[elbert]

Change windows power settings to performance from balance.

Are all the games valve installs copied from another system? If so go into games library and right click to select properties. Then select launch command and type -autoconfig. This will force the games to reaquire your hardware settings. May not help but worth a shot.

 

[bigpinkdragon286]

So far, it would seem to me that you've been testing your CPU and GPU separately, while games tend to load both.

I would use Furmark to stress the GPU, along with either Intel Burn Test or the Furmark CPU burner to stress the CPU. Intel Burn Test gives me a few more °C on the CPU while it's running, but the CPU burner built into Furmark tends to leave enough resources that the GPU can continuously be stressed to 99%.

There is some logic in wanting both CPU and GPU not to be 100% pegged continuously, as in the case of most computing scenarios such as gaming, loads vary quite dramatically, and can do so rather quickly. It can be in the transitions to or from a high load that you experience that momentary instability. That is why Intel Burn Test is cyclical. It makes certain to let the CPU load go up and down, which in turn forces the GPU load to go up and down as well, when run together. You just need to figure out what settings to use with Intel Burn Test that allows the highest simultaneous GPU load with Furmark.

The usual caveats apply. Furmark can and has caused GPU power delivery circuits to be overwhelmed and damaged as a result of running it. I'm sure the same is possible with Intel Burn Test, just haven't read of cases where it happened. This is usually because the circuit doesn't have enough capacity, and running full tilt was actually never intended by the manufacturer. It's not because either of these tests can intentionally destroy the equipment, it's because the equipment can't handle the stress of running itself at wide open throttle. Make sure you keep an eye on temperatures and voltage lines. It's up to you to make sure nothing gets out of hand and self-destructs.

 

[Amet Monegro]

Disarm your PC and check Motherboard and CPU, maybe has some bend pins, put protection between motherboard and case

 

[TechnoStyle]

Change windows power settings to performance from balance.

Are all the games valve installs copied from another system? If so go into games library and right click to select properties. Then select launch command and type -autoconfig. This will force the games to reaquire your hardware settings. May not help but worth a shot.

Power settings didn't change the results.
All components are new. Windows 10 is fresh installed. Also games, steam etc. fresh/ new.

 

[TechnoStyle]

Today, I ran 3DMark Demo and PC crashed immediately.
I did a lot of test for hardware and tried a lot of suggestion about software.
I'm absolutely sure that my GPU is defective.
I'll RMA my graphics card. Thank you for all help. Much appreciated.

 

[Amet Monegro]

Change windows power settings to performance from balance.

Are all the games valve installs copied from another system? If so go into games library and right click to select properties. Then select launch command and type -autoconfig. This will force the games to reaquire your hardware settings. May not help but worth a shot.

Power settings didn't change the results.
All components are new. Windows 10 is fresh installed. Also games, steam etc. fresh/ new.

Check i said

 

[sniper360]

Do you have a spare PSU to try out? Or borrow one to test?
Even though you have a reputable PSU, it may be worth checking out.
I have had an electrician with calibrated equipment check out one of my older PSUs...it should not have been rated GOLD for sure.
Sounds like you've performed the logical steps already...

 

[elbert]

If your overclocking reset your bios and test at stock. Remove/uninstall any 3rd parts programs such as ai suit, ryzen master, MSI afterburner, and any others. You have tested all your hardware so has to be malfunctioning software. You maybe should also try reinstall, update OS, and update drivers then reinstall games and nothing else. Test your games on a clean install. The only wild card is your PSU so if you swap it out with a know good PSU to test with may help.

 

[Zachary Singer-Englar]

Have you double checked that the power cables to your Mobo (24 and 8 pin) and to your GPU are fully seated? Have you ensured that the PC is set to prefer the RX480 over onboard in BIOS? Please also confirm that you are plugged in to the RX480. You may want to reinstall the graphic card driver in safe mode and run whatever the manufacturer provides as far as a cleanup application goes to remove any old driver data first.

It's possible you have a defective part. You can test further to confirm that the issue is not with your CPU by removing the GPU and running some light gaming load tests with onboard video (if your mobo has onboard vdeo), though clearly not at the same settings. Your motherboard manufacturer might also have some diagnostic tools available on their website or built in that you can boot to that might help.

 

[AniChatt]

I am also facing somewhat similar problem like yours. And I dam sure it is not my GPU or PSU. I think something in the windows causing some conflict. here is my problem summary,
1. On day one after I installed all my components I have ensured the cpu is sitting properly before closing the latch. Power on my system of old Windwos 10 ssd and before going to windows I setup bios with with latest Asus 0805 AGESA 1.0.0.6a. That time everything was in pure default and my ram profile was auto at 2400MHz 1.20v.
2. Then clean installed Anniversary edition of Win10 and installed all latest drivers available and windows update (not founders) it is the basic cumulative updates.
3. Started to experience RAM issue with mobo amber light blinking sometimes as well as VGA error led sometimes the the display turns off.
4. Then I started to play with RAM speed (not timing: it is set by the mobo) and found the situation is much stable if I turn of the default D.O.C.P profile which by default switches on 3200Mhz profile (14 timing as written on the stick 1.35v).
5. Again experienced the problem. Sometimes VGA light is on or the RAM light is on.
6. After few days no error led but display of, PC freeze even power button is not working
7. I downloaded Windows 1703 founders clean iso and installed it fresh. The pc was working fantastic that time for up to 7hrs. Completed all installation and update with asus AISuite III till this time.
8. On next day morning after I woke up and started my pc and the display gone in less than 5 mins. Restarted by switching off UPS. Restarted again same problem and it continued.
9. I have checked it with different bios setting like> default, SMT on/off, LLC Auto/Regular/medium, SoC LLC Auto/Regular, Slight overclocking without changing voltage, Turned CPB off, even tried under-volting my CPU upto -0.03125 offset.
I observed one strange behavior of this system: In default when CPB is on this mobo pushes crazy voltage spikes like 1.52v as shown in the left corner of bios but immediately switches back to 1.21v.
One more thing, If I increase the core ratio say 38 then the voltage automatically increases no matter what I selected in bios. e.g. 38 ratio and core voltage is 1.375(bios) - 0.03125 (offset) but during the stress test HW info and all other programme showing the voltage reaching upto 1.40v on core (LLC of Core and SoC both was regular). But the problem of display is there for every single experiment and it is random.
10. Again clean installed Windows 10 1703 and this time no AISuite, fan controlled by bios, chipset driver of AMD not asus, tested with two different GPU i.e. GTX970 and R9-290 (cleaned driver as well). Nothing fruitful.
11. Also tested the memory with memtest for total of 14 Hours and found Zero issues. Also tested with few of my different PSU cables just to ensure it.

Summary:
This problem is absolute random. Sometimes it takes 5 mins and sometimes 7 hr. But it is there. Nothing to do with overclock, it remains exactly same. Situation improves with 3200Mhz ram clock + R9-290 compared to Auto RAM frequency + GTX970. Found one issue in Prime 95, one worker stopped and returned fatal error, but as per G.Skill tech support it may occur rarely and not a ram problem. Most of the time when it occurs, on the mobo any diagnostic light not glows. When I ac power off the system of course there will be some power in the mobo, now unless it is drained completely, if I switch on the ac power the power button fails to turn on the PC. PC only turns on after the power is drained. But in bios it never ever shows any error or failure or press F1 to config BIOS. But it happens (error and bios set up needed) when the RAM failure occurs with amber mobo led on/blinking but that time, power button works perfectly to switch off the PC (force shut down).
So the problem is too complicated for me.

Not fixed yet and waiting for a new bios. After checking with new bios I will RMA my motherboard from Asus.
Thread: https://community.amd.com/thread/218356

 

[TechnoStyle]

Update:
I RMA' ed my graphics card and service found that my GPU is defective. I refund my money back.

 

[TechnoStyle]

Last Update: Here is the result: I took my money back and bought ASUS Dual GTX 1060. Problem solved. GPU of my Strix graphics card was defective.
Don't judge me, there is no RX series graphics card because of the miners.