PC Drivers & Software

[agoosh]

Specifications:

OS - Windows 10 Pro

Processor - Intel Core i5 4590 3.30GHz

Motherboard - ASUS H81M-P Plus

GPU - AMD Radeon RX 570

Power supply - Corsair CX550

After first building the computer, everything worked fine. Then after about a week the graphics card wouldn't work so I exchanged it for a new (bought used) one. The AMD driver could not detect the GPU, and a friend IT technical put the GPU card in a different computer and it was running fine, so we thought it was the motherboard.

I got a new motherboard and the same issue remains - the card is in properly, it has power as the blue "sapphire" light is on, but it is not detected. It does not come up in BIOS, even when I changed the setting to use the GPU, and it does not show up in the device manager.

After going on several forums here is a list of the things I have done to try fix the issue:

- updated BIOS from the 0501 version to the 3602 version (29/4/2018)

- updated Windows

- installed newest Microsoft Visual C++ and uninstalled all old ones

- ran a driver cleaning software (the AMD one)

- ran DISM (showed no corruption)

- ran SFC (showed no errors)

Has anyone got any ideas as to what else I can try? I have no idea why it stopped working and now won't work at all. I checked and all parts are compatible with each other, and after doing everything suggested online to detect the hardware it is still not working.

Thank you for any suggestions.

[elstaci]

I see that your Asus Motherboard only has one PCIe x16 slot.

Do you have both GPU Power connectors connected? 8 pin plus the optional 6 pin power on the GPU Card?

Do you have your monitor cables connected to the Motherboard (Intel integrated graphics) or the Rx 570 (Discrete Graphics)?

If you card worked fine in another computer, then it is possible you have a hardware failure or failing in your computer setup.

Download a free diagnostic program called OCCT. Run the PSU test on it and see what type of results you are getting. After a test is run or is terminated early, it still creates a OCCT Folder inside of Windows "Documents". Check the following graphic images:

PSU 3.3 VDC output, 5.0 VDC output, and especially the 12 VDC output.  See that they are all within tolerances for each Voltage output.

It is possible your PSU is not supply proper Voltage to the GPU card thus not making known to BIOS or Windows.

Sapphire Support has its own AMD Drivers to download. Try those and see it your GPU is recognized. Download it from here (for 8 gbyte Card): Request Rejected

Here is the latest AMD Driver for the RX 570 (18.9.2) : Radeon™ RX 570 Drivers & Support | AMD

Follow this basic procedure in installing AMD Drivers on your Computer:

install AMD driver:

1) Download the correct AMD Full Set of drivers from AMD Support. Make sure your Windows is fully updated via Windows Update. Windows Must be fully updated because the latest AMD Drivers requires all the latest "Optional" and "Recommended" updates to be installed.

2) Use Windows Uninstall to uninstall current AMD driver and software and disconnect the internet from your computer. Then use DDU (Display Driver Uninstaller) Display Driver Uninstaller (DDU) from Wagnardsoft  Forum in Safe Mode. This will eliminate all traces of the AMD driver and software from your computer.

Delete C: \ AMD folder from the Root Directory. Reboot

3) Go to Device Manager and click "Display Adapter" and make sure you are on the MS Basic Display Adapter. If not, uninstall the AMD driver using Properties.

4) Try reinstalling the AMD FULL SET OF DRIVERS that you have downloaded manually. Make sure you disable the Internet to prevent Windows from installing a newer version. So configure windows to prevent it from updating drivers via windows update. So it has been mentioned to disable any anti-virus programs before installing AMD Drivers.

5) If the new AMD drivers installs and works correctly, delete again the C: \ AMD folder from the root directory. To save space on the HDD.

6) Enable both the Internet and Anti-Virus program (if applicable).

7) Go back to Device Manager and check your GPU card driver is working and identified correctly.

Note: the link says "Request Rejected" but when I click on the link it went to the correct Webpage.

[agoosh]

Hi, thank you for your reply.

I have downloaded the OCCT program as suggested, and ran the test. On

Infinite, I was not getting any results, and on automatic, the test stops

after the "idle period". I have a lot of results and I do not know which

ones are the VDC one you've said to check. Could this be a power problem?

I'd really appreciate your help.

Thank you.

[agoosh]

Here is what came up after a 5 minute test

[elstaci]

It is almost impossible to read the graphs you posted. It is too small and blurry.

At the beginning and end of each OCCT test to allow the hardware to go back to normal temperatures and state. After it idles for about a minute, OCCT then runs the actual test for the hardware being tested (CPU or GPU or PSU).

From what I saw it doesn't mention the 3.3 vdc , 5.0 vdc, nor the 12 vdc voltages. Which test did you run? What size is your PSU?  Wattage Rating?

Let the PSU Test run for a full hour and then upload the results. just keep an eye on it. The PSU test has been known to damage weak or failing or cheaply made PSU from what I have read online. Otherwise it should run normally.

When running the PSU Test you should see constant data concerning the various PSU voltage outputs and temperatures of GPU and CPU.

Also run the CPU test to see if that checks good.

This may not be a hardware issue but when you said that even BIOS doesn't see your card, it pretty much eliminates the AMD driver since they are loaded when Windows loads up. I am just eliminating the PSU from not supplying enough power to the GPU card.

Also make sure both GPU Power connectors are connected to the card directly from the PSU. If one is missing (6 Pin) it might be the reason why BIOS doesn't detect it even though all the GPU lights and fans work.

This is an old Tom's Hardware thread about plugging in both GPU Power connectors directly from the PSU: Power connectors on my video card - [Solved] - Systems  But it still is valid for today's newer GPU Cards.

[elstaci]

This person had almost identical problem as yours in where his GPU was good on another computer but not his. BIOS even wouldn't detect it. BIOS not detecting GPU - [Solved] - Graphics Cards​ . Turn out to be a GPU alignment problem on motherboard Slot or a bad power connection.

This thread from Asus with a similar problem suggest it could be a faulty CPU.: HELP - BIOS not detecting GPU - Maximus ix hero .

That is a good reason to run OCCT CPU Test and see what results you get

But yours could also be a power issue or the power cables to the GPU card are bad, bad CPU or alignment issue or even a BIOS not being configured correctly. That is why I am trying to eliminate the PSU first.

[agoosh]

Sorry for it being small, I just posted it to show you I couldn't find the 3.3

vdc , 5.0 vdc, nor the 12 vdc voltages.

Here's the power supply info:

I ran the PSU test, it stops at 59s once the idle stops. I left it alone

for 2h and it did not go past it, here's what it looks like when it stops:

It stopped creating reports in the OCCT file as well.

The CPU test stopped, as Core #4 was over maximum value:

Also, the GPU has all the power connected directly from the power supply,

both the 8 pin and the 6 pin.

Thank you so much for your help.

[elstaci]

I googled your error for the CPU test and it indicates your CPU overheated.

The image you posted shows that you stopped the PSU test yourself. It says User Cancelled.

DId you actually stop the test or did the test abruptly stopped by itself?

Also it is important that you run all OCCT Test in "FULL SCREEN" for maximum stress testing. I noticed that on the PSU test it was not clicked.

The PSU Test is done by OCCT running both the CPU & GPU test at the same time which put maximum load on the PSU. If you let the PSU test run on Automatic until it stops by itself or finishes. If it stops by itself within a couple of minutes then it could be again due to your CPU overheating.

Before running the CPU tests again, Reset your Motherboard BIOS to "Default" or best optimal settings. One person found out BIOS had his CPU on Turbo Mode which caused it to overheat when tested by OCCT.

Run the CPU Test again.  When running the CPU tests keep a very close eye on the temperatures. Don't run the CPU -Linpack since OCCT website says the first test is the best one to run. The one you ran originally.

To be safe, leave one side of your computer case open in case it is due to poor air circulation and to see if the CPU fan is spinning correctly.

Before running any more tests, delete the OCCT Folder in Documents.

Post back your results again after checking your CPU Cooler and that the CPU Fan is working.

Need to take a look at your CPU and make sure the CPU Cooler is attached correctly to the CPU and the Fan is working. Mention this to your IT Friend and maybe he can check out your CPU on his computer if his motherboard uses the same CPU Socket or he can take a look at your Intel CPU and see if he sees anything out of the ordinary with it.

[agoosh]

The fan works normally, speeds up when more things are open etc so I'm not

sure why it let the CPU overheat so easily. Other than the OCCT, I don't

have anything opened.

I didn't stop the power supply test, it went into the "user cancelled" by

itself, I just pressed ON and let it do it's thing.

Here are results of the CPU: Linpack test, which once again stopped itself.

I hope the graphs will be big enough this time. It took 1 minute 19 seconds

to overheat, the fan kicked in and it took longer this time for the

temperature to go back down (about 5-10 minutes).

The results for the temperature for core 0, 2, and 6 were similar to those

of 4.

Sadly I won't see the technical for a few weeks now which is why I was

hoping to maybe find the problem myself. Would a different/better CPU fan

change anything at all?

Thank you for your help.

[elstaci]

Maybe OCCT is misreading the CPU Temperature.

Download HWMonitor and run it while playing a game or running a program that will cause the CPU to run at 60 to 100% load. Check the temperature from HWMonitor and see if the CPU temperatures are high.  See if it is similar to the OCCT CPU temperatures. Also don't run the CPU-LINPACK test. This produces the most heat. Just use the OCCT-CPU test only with SMALL DATA PACK. This will generate the most heat using this test.

I run OCCT on my FX 8350 without any overheating problem and it does show the proper temperature compared to the other monitoring software I have running.

I don't know if the CPU Overheating may be causing your GPU card not to be noticed. But it might indicate a problem with your CPU cooling.

If OCCT CPU test is showing that your CPU is overheating that is probably the reason why the PSU Test also stopped by itself. Like I mentioned, The PSU test involves running the CPU-Linpack and GPU 3d test at the same time producing the heaviest load on the PSU.

[elstaci]

I just ran the OCCT-CPU test with Small Packet for 4 minutes. Here are the results. The CPU Temperatures were identical to my Hardware monitoring software. I also included the Graphics that show my PSU Outputs of 3.3, 5.0, and 12 VDC:

[elstaci]

I noticed that on your screen shots of OCCT the Monitoring part which important is not being shown. You need to click on the RIght side of the menu where it says Monitoring to show that window. That is the one the shows all the pertinent information when you were running the test in a nutshell. It show Max and Min temperatures and voltages of the CPU. GPU, and PSU outputs.

Just ran the OCCT PSU Test for about 5 minutes. Here are the results:

If you noticed all my PSU Outputs are within tolerances during the test. The hottest my CPU got was 56 c. But then again I am using a very good CPU Cooler with dual fans on the heatsink.

[agoosh]

As embarrassing as it sounds, I checked and the CPU fan wasn't sitting in

correctly. I got frustrated and took everything apart and put it together,

and now the fan is sitting flush to the CPU. Also changed the fan setting

in BIOS from standard to turbo (duh, it should have been like that before)

I ran the CPU: OCCT test again, and got much, much better results. I

stopped it at 30 minutes as the temperatures were staying at the same range:

Ran the test again, as you suggested, on Small Data Set, for another 30

minutes:

At least that's one thing sorted. Tried to install the Radeon Software

again, with no luck - "AMD installer cannot continue since AMD graphics

hardware was not found". Using the DVD that came with the GPU to install

drivers was not successful either.

Thank you.

[elstaci]

Okay, I see a problem. OCCT is not showing the 5.0 VDC nor the 12 VDC from your PSU.

Edit: Found it that some Hardware monitoring programs including OCCT may not be able to read the Motherboard's sensors. Thus the 12 and 5 Vdc doesn't show up.

Click on the upper right corner on the far left square until you get this view. I am hoping it will show the 12 and 5 vdc like on mine.

Download and install the Free version of HWMonitor and see what the PSU 5.0 and 12 vdc are. Especially under heavy loads. You can probably run OCCT PSU or GPU test in Windowed view so that you can see the Outputs from the PSU while watching HWMonitor.

See if the 12 VDC is below 11.4 VDC under heavy loads.

Well glad you were able to find and solve a critical CPU installation problem. Now lets see if the PSU is your problem.

[elstaci]

Edit: Deleted comment about installing ASUS AI SUITE III. kingfish​ made a good point in mentioning about it might conflict with AMD Driver software.

Forgot that 3rd Party controlling software does at time interfere or have conflicts with AMD Driver APPs.

In that case, If HWMonitor doesn't show the PSU Outputs then try this one that many Users seems to recommend: https://www.hwinfo.com/download.php . Hopefully one of the two will show the motherboard's PSU 12 and 5 volts outputs.

[kingfish]

AI suite and other third party controllers conflict with AMD settings and Wattman controls. It's best to uninstall it.

[angeh]

what???????????? but i control my case fans and my cpu with Al Suite 3..... i recently got a rx 570 which i have not yet installed it on my board, and now you are telling me this :/.. sigh i think i made a mistake by getting a radeon card /facepalm.

[kingfish]

You use one or the other. If your going to use the Asus control..do not use the Wattman controls.

[angeh]

okay now that you mention that, is it super recommended to use wattman for a radeon card for control/undervolting/etc or can i just keep my system with the al suite 3 for my cpu/case fans(intel cpu ) and for the gpu card msi afterburner? (which is what im using right now but with a gtx 750 ti)

[kingfish]

I don't think you understand. Any third party controller..these include Asus and Afterburner. Do what you want..if it doesn't work correctly get rid of them.

[kingfish]

Since AMD is new to you, I suggest you read this before your next question(s) > Radeon™ Software Help Center | AMD

[kingfish]

I use Open hardware Monitor...here's what it shows:

[kingfish]

Maybe CPU-Z?

Download CPU-Z | TechPowerUp

[agoosh]

Thank you for your replies. Nothing seems to fix the card not being

detected, and I have now started to consider giving this one back and

buying a different one for the refunded money. A shop with used computer

parts near me is selling am NVIDIA GTX 1060 3GB for a bit more than this

one was, which would be a good fix for me if that would work haha. I've

spent so much time trying to fix this set up that it's beginning to really

annoy me.

Thanks

[elstaci]

Sad to hear, but I don't blame you. It probably is a defective card.

What threw me off was that you said it worked fine in another computer.

Normally you are taking a risk purchasing a Used hardware like GPU Cards. You never know what type of abuse or reasons why it was replaced the first time. But if you are on a tight budget, then buying a used hardware is the next best thing as long as you have some sort of Warranty on it. At least 30 to 90 days.

Good luck with your future GPU card.

[agoosh]

Got the new card, works perfect. The shop tested the old card and it passed

all their criteria so I really have no idea why it didn't work for me.

The card I have now has a 1 year warranty which is decent, as a student I

sadly can't afford a new GPU.

On a side note, I checked in BIOS and I got the following readings:

3.3V - at 3.296V

5V - at 4.960V

12V - at 12.096V

Anyway, thank you very much for all yours help

[elstaci]

Your PSU Outputs are great.

Outstanding getting a one year Warranty on a used GPU card (Refurbished).

LoL, Your Asus Motherboard is Anti-AMD

[kahless_uk]

It's a Sapphire card right? have you checked for a bios selector switch?

iirc Sapphire cards need to have their selector switch match the boot settings in your bios, legacy mode for bios and uefi mode for uefi (unless your uefi is set to legacy boot) which "might" explain why it works in one machine and then not in another