Problems with PC after upgrading gpu and psu! MSI R9 290! Need a Miricle!:(

[nathannnn22]

Right, so I'll start at the beginning... I got a new gpu. MSI R9 290 4GB Gaming OC edition, and a new psu, Coolermaster G750M 750w, for christmas. I installed them correctly on christmas morning and got it up and running absolutely fine. It ran for about 4 hours, and then all of a sudden my screen went all black, and my PC reebooted; I thought it had just overheated, but it hadn't, the temperatures were all fine. And ever since then, my PC now randomly reebots whenever I have the 290 in. Here are my current PC specs:

ASUS M5a78l-m/ USB 3 mobo
AMD fx 6300 cpu
8GB generic ddr3 ram (i think)
1tb HDD
MSI r9 290 (new)
Coolermaster 750w psu (new)

I tried putting my old gpu back in, ASUS GTX 650, and my pc runs fine, so it can't be the new psu. I've uninstalled all drivers for AMD and Nvidia and still doesnt help, i've enabled force constant voltage with msi afterburner, I've updated my bios, and ive put my old 430w psu back in to power the system, and used just the 750w psu to power only the gpu... After all that I decided to return the card, however they came back to me and said that they couldn't find any fault with the card? I really don't think they tested it properly... However, like right now, i just received it back so i put it back in, still blackscreens, but i left it on the desktop for ages, surfed the web, wrote this, and it hasn't blackscreened at all? It has blackscreened before when I was just surfing the web and on the desktop? Please can someone help me, I honestly don't know what else to do... Thank you so much for reading!

 

[bgunner]

Have you run a driver sweeper program, such as Driver Fusion, to completly remove all Display Drivers to ensure a clean install of the new GPU driver?

Have you tried different driver versions?

After updating the Bios did you stress test the the system with the Older GPU installed? Doing so will ensure the new PSU and the rest of the system is stable before adding in the component that you believe is the culprit. I'm thinking you may need to up the CPU-NB voltage but testing will need to be done to be sure this is the case.

 

[nathannnn22]

Have you run a driver sweeper program, such as Driver Fusion, to completly remove all Display Drivers to ensure a clean install of the new GPU driver?

Have you tried different driver versions?

After updating the Bios did you stress test the the system with the Older GPU installed? Doing so will ensure the new PSU and the rest of the system is stable before adding in the component that you believe is the culprit. I'm thinking you may need to up the CPU-NB voltage but testing will need to be done to be sure this is the case.

Yes I used display driver uninstaller. Not different driver versions no, only used latest, and the beta ones. I didn't stress test it but i played some fairly demanding games whilst waiting for my 290 to be returned. In the bios i did change the PCIe bus settings from auto to 110 and the chipset overvault to 1.3 if thats what you mean?

 

[bgunner]

The PCI-e should be left at 100. Adding speed to this can cause instability. I recommend setting it to 100 till you figure out where the issue is coming from.

Please have a temp monitoring program before doing the following.

1. test the system with Memtest86+ for 1 pass if using ver. 5.01 or over and 11 passes if using ver. 4.3. You will need to put this on a flash drive or CD and needs to run as the boot device. This will test the ram to ensure the ram is stable at the settings you have them set at. IF ram is unstable it can show up as other things.

2. Test teh system with Prime95 for 2 hours minimum. This program is designed to stress the CPU to find out if it is stable. * Warning * watch your temps while running this program as it will heat up and could overheat if your cooling is not enough.

3. test the system with Intel Burn Test. Run this program set to maximum stress and for 20 runs. This program is designed to stress the CPU to find out if it is stable. * Warning * watch your temps while running this program as it will heat up and could overheat if your cooling is not enough.

4. (With older GPU installed) test the system with OCCT set to GPU test and error Checking enabled for 1 hour. This program will stress the GPU and help stress the North Bridge to show any instability in this area.

IF any program errors it will help point us to the solution. If no programs error, then we know the system is stable and the problem lies with the GPU, driver or the GPU being incompatible with the motherboard. Although this last option is rare it can not be discounted till all other avenues have been explored.

 

[nathannnn22]

The PCI-e should be left at 100. Adding speed to this can cause instability. I recommend setting it to 100 till you figure out where the issue is coming from.

Please have a temp monitoring program before doing the following.

1. test the system with Memtest86+ for 1 pass if using ver. 5.01 or over and 11 passes if using ver. 4.3. You will need to put this on a flash drive or CD and needs to run as the boot device. This will test the ram to ensure the ram is stable at the settings you have them set at. IF ram is unstable it can show up as other things.

2. Test teh system with Prime95 for 2 hours minimum. This program is designed to stress the CPU to find out if it is stable. * Warning * watch your temps while running this program as it will heat up and could overheat if your cooling is not enough.

3. test the system with Intel Burn Test. Run this program set to maximum stress and for 20 runs. This program is designed to stress the CPU to find out if it is stable. * Warning * watch your temps while running this program as it will heat up and could overheat if your cooling is not enough.

4. (With older GPU installed) test the system with OCCT set to GPU test and error Checking enabled for 1 hour. This program will stress the GPU and help stress the North Bridge to show any instability in this area.

IF any program errors it will help point us to the solution. If no programs error, then we know the system is stable and the problem lies with the GPU, driver or the GPU being incompatible with the motherboard. Although this last option is rare it can not be discounted till all other avenues have been explored.

Okay I will try them. Are the first three with the 290? What shall I do if my PC blackscreens and reboots when performing the test? Shall I perform #4 with the new gpu too? also you told me to use Intel Burn test, my cpu is AMD will I still be able to use the program? Thanks! Am taking the 290 to a friend's tomorrow who has the exact same card and seeing if it works in his PC. Will get back to you. Thanks.

 

[bgunner]

All tests should be done with the Old GPU.

If the PC black screens and reboots consider that test a fail.

we are trying to start off with a stable machine. THis means having the parts in it that are known to be stable. These tests are designed to stress teh system past what gaming can put them through and will find errors or issues you may have missed.

Yes it is a good Idea to try your card in a different machine to see it the issue remains. Good troubleshooting on your part.

I can not stress enough how important it is to stress test the system with the Old GPU first. Running all tests with the old GPU gives you a base line to start from. IF the system is stable for all tests with the Old GPU but only fails the GPU test with the new one we know the issue lay only with the GPU, drivers or compatibility. We would know this because the system was tested stable before inserting the New GPU. BUT if the system is unstable with the Old GPU you might have just found the issue that is causing the crashing. Testing needs to be done in a certain order to ensure the test results are accurate.

The only test that does not need to be redone is the Memtest86+. Once this test is done it will not change the results having a different GPU installed.

so in short:
with old GPU run all 4 tests.

with new GPU run tests 2,3 and 4. IF any error or the system crashes consider the test a failed test.

 

[nathannnn22]

All tests should be done with the Old GPU.

If the PC black screens and reboots consider that test a fail.

we are trying to start off with a stable machine. THis means having the parts in it that are known to be stable. These tests are designed to stress teh system past what gaming can put them through and will find errors or issues you may have missed.

Yes it is a good Idea to try your card in a different machine to see it the issue remains. Good troubleshooting on your part.

I can not stress enough how important it is to stress test the system with the Old GPU first. Running all tests with the old GPU gives you a base line to start from. IF the system is stable for all tests with the Old GPU but only fails the GPU test with the new one we know the issue lay only with the GPU, drivers or compatibility. We would know this because the system was tested stable before inserting the New GPU. BUT if the system is unstable with the Old GPU you might have just found the issue that is causing the crashing. Testing needs to be done in a certain order to ensure the test results are accurate.

The only test that does not need to be redone is the Memtest86+. Once this test is done it will not change the results having a different GPU installed.

so in short:
with old GPU run all 4 tests.

with new GPU run tests 2,3 and 4. IF any error or the system crashes consider the test a failed test.

Hi, I have completed all tests and with the gtx 650 I was able to complete all of them, however with the 290 I wasn't able to complete them as my PC kept rebooting. I have already returned the card once and they said that it worked fine in their PCs. I have now returned it a second time and have asked them to conduct more tests. I honestly think there's something wrong with the card, surely if it wasn't compatible, it wouldnt have run stable for 4 hours when I first installed it? Thanks

 

[bgunner]

OK. I just need some clarification. When you uninstall the drivers you use a completely different program, other than the uninstall option built into the Driver installer, to uninstall the drivers and programs? I can not stress enough that the uninstall options from the installers can leave behind drivers that can mess up a new installation mostly when going from AMD to Nvidia or vise versa.

IF you have used a driver sweeping program to completely remove all of the old and current drivers and THe company sends your GPU back the only explanation is your Motherboard and that particular GPU are incompatible. I have run a crossed this only a few times But it does happen. The Motherboard just does not like that exact MSI model. Now if you swap brands to say an Sapphire, Asus, XFX, Power color, Gigabyte they would would work fine. The lasts one I ran across like this was a Asus R9 280 and a Dell Motherboard. The GPU tested fine in every other system But that one. I have tried Different PSU's, Ram, CPU, HDD's, Cases and finally I tried a more powerful GPU on the Dell board. The More powerful GPU worked fine in the original Dell setup. It came down to the GPU not being compatible with that motherboard. This is rare but does happen. Usually the issue is no display at all but as long as the system tests good in all area's and the proper Hardware requirements are there the last option is this.

Before saying this is the case be sure ALL older drivers have been completely uninstalled. The only sure fire Way i have found to do so is by using a separate driver sweeping program as it removes current and previously installed drivers that get used as backup's/roll backs.