Windows 10: Several different BSOD errors

Ran sfc, had no integrity violations.

I have tested my RAM sticks and lo and behold, one is defective. I have had my graphics shipped to California for a warranty repair/replacement... here's to hoping that it was defective too and they replace it with a 8gb RX 480

 

that's good news. I hope that solves your problems. Good job on the graphics card. I would seriously suspect the ram. Please let us know how it comes out.

 

OK so I have replaced the RAM sticks and and that seems to be good. I received a refurbished R9390 from the manufacturer.

The previous 390 didn't click in 100% into the mobo due to the heatsink being in the way. Is the fact the 390 doesn't "click" in 100% detrimental to the cards performance? If it's not I will likely have to buy a better PSU to handle its demands.

In the alternative, I can keep the 390 in the box and try to sell it for a better price. The problem is that the RX 480 is very close to the 390 in performance and it is cheaper brand new. I've only gotten 1 low ball offer for it about 200 dollars under asking, but I'm worried no one will want this card in a couple of months.

 

I really can't answer about the clicking in . The clicking is a retainer to hold it into place and make sure it doesn't vibrate loose. If the card made complete contact in the slot, there should be no difference in power. I haven't looked it up, but I would suspect your power supply is the minimum recommended for the card. If you do buy another power supply, stick with Seasonic. They are great power supplies, like Corsair AX and HX models, XFX and a few other brands. Seasonic is generally thought of as the best and make some of the other better brand power supplies.

Nvidia just came out with new cards. So, I would suspect AMD will come out with new ones soon to compete with Nvidia. That would make your current cards worth less as they would be a generation older.

 

Here's the update:

• I sold the 390 since I wanted to get a GPU that works with my 550 power supply.
• As noted, I was still getting memory errors without the 390 installed; bought new RAM and I was able to play Overwatch on low settings with the DirectHD graphics on the mobo.
• Fast forward to today and I bought a brand new RX480. I am now getting atikmpag.sys errors when I play Rocket League. I played a round, and when I pressed "Exit game", it went into a boot cycle where the system would restart, load windows 10, then crash and reset.
• I am confident this brand new card isn't the problem. I will get someone to test it on a new system just in case.
• My guess is either the mobo, PSU, or CPU are the culprit. I will try to do the CPU test again and see what results I get. Is there any way I can test my mobo and PSU?

 

The best way to test a power supply is with a digital multimeter. Make sure it is digital. There is no way to test a motherboard without a lot of special equipment that only the manufacturers have. But, go into bios and tell me the values of the +12V, +5V and +3.3V.

 

your card's recommended power supply is 500W. Which is at your PSU's max output. I'm not saying it is your PSU, but it is the graphics drivers that are crashing. Let me suggest, Download the latest driver from AMD (Download Drivers) uninstall your current drivers. then follow this tutorial from sevenforums to remove all leftovers http://www.sevenforums.com/tutorials...r-how-use.html. It works the same with windows 10. Just choose AMD and nothing else. when you reboot, install your new driver that you downloaded from AMD, but select Custom Install, only install the Graphics driver, nothing else tht you don't have to. Do not select the express install. See if that works any better for you.

 

I tried the driver suggestion. I wiped it with DDU, then installed just the drivers from AMD - newest from July 19 2016. After the restart the computer went into another boot cycle and produced more driver crashes...
I tested the card on my friend's computer, and it was able to boot up the system no problem, and was able to load rocket league and witcher opening screens without issue.
I did the Prime 95 Test. The CPU did max pretty quickly to 95°C-100°C. Nonetheless I kept it going for about 3 hours 5 minutes without any crashes.

I will check my mobo voltage, and see if there is a BIOS update for the mobo since May 17/16.

 

+ 12.00v : +12.192 v
+ 5.00v: +5.160 v
+ 3.30v: 3.424 v

 

Your voltages look OK. It seems that your CPU was throttling during the Prime tests. That is entirely too hot for your CPU.

Could you just try putting your card in a different PCIe slot? See if that helps any.

 

My case is too small, the PSU blocks the other PCIe slot and i physically can't reach it. Should I possibly reapply the thermal paste to the CPU and reseat the fan?

FYI I only have the stock fan on the CPU and the case only comes with one back fan. My apartment has been quite hot recently, whereas when I tested it in the winter/spring the room was colder.

 

If you are going to run Prime95, you are going to have to get a 3RD party CPU Cooler. The Intel cooler just isn't capable of handling that much heat. Have you considered doing a clean install? If you have a spare hard drive, we could do the clean install on that and save your current install, if it turns out not to help.

 

I wasn't aware, the other person (Arc) said to use Prime 95 to test the CPU.
I can potentially try a clean install on my regular hard-drive in the system and not the SSD.

 

I understand. Prime95 creates a massive amount of heat and the stock Intel cooler just is not capable of handling that amount of heat. But, you've run it now and unless you had a BSOD, lockup or other unusual activities, that part is done. If I understood correct, you put your card in another machine and it ran OK. But, you didn't really run it hard to see if it had any problems. In my opinion, you have a problem with the card, drivers, motherboard or PSU. If you are satisfied that the card itself is OK, then the problem lies in one of the other areas. There is no test for the motherboard. The PSU can only be checked with any certainty with a digital Multimeter. However, the values you gave from the BIOS are good. But, that only tells us what they are at idle, not what they do when under stress. You have a very good quality PSU and I would expect it to be ok, but even the best ones malfunction on occasion. There is also the possibility that you have another program installed that conflict with the graphics drivers. That's where we stand right now, as I see it.

Have you run driver verifier? We could try that first. In fact, even if we have done it, let's try it again. This will show you how. If you have any questions, feel free to ask. Please run it for 24 hours and a maximum of 48 hours. It places stress on the drivers and tries to make the computer crash in hopes of catching the bad driver. Just enable it on non Microsoft drivers. Use it anyway you want to. If there is something you do that makes it crash, do that often. Then upload the BSOD posting instructions again. This is the tutorial the video is based on. https://www.tenforums.com/tutorials/5...dows-10-a.html

[youtube]BjsofLPNosc[/youtube]

 

I agree with your statements overall. My friend didn't want to uninstall his other AMD drivers for this card so we didn't test it extensively, but it seems like the GPU is fine.

My CPU did not crash so I think it definitely is in working order. It is possible the PSU is faulty, as when I add the GPU the computer seems to hang at pretty much any point. Without it the system works fine.

I did do Driver Verifier before, and it preemptively confirmed the memory/RAM issue. I can try it again but the drivers seem to crash the computer even before it fully boots up, then it fails to boot another 2-3 times just like with the other card. Or I boot up a game and play rocket league for about 5 minutes before crashing. This is what leads me to believe the mobo is the likely culprit, as the issues haven't changed even with a new card.

I will try DV first and post the results. After that I can try to reinstall W10 on my HDD to rule out the SSD, although I have a feeling this isn't the issue.