RX 480 8GB upgrade problems

DarraghR

I have upgraded my GPU from a R7 290 to a RX 480 8GB Sapphire Nitro edition. I first encountered problems when I began to run DOOM at Ultra settings. Upon doing so it would power down my computer and restart. I ran Mortal Kombat X with full settings prior to this.

So I thought maybe update some drivers and software. Next i ran an older game (Dead Space 2) but this crashes after 5 seconds of the title screen...
I have used AMD Clean Uninstall Utility and went through ADD/REMOVE software on Windows 10 but still have crashing.

Below is a blue screen that has popped up a couple of times now



Updating my BIOS was also mentioned but there are no updates for my system on HPs website under Windows 10. There is for Windows 8.1 but they will not install.

If someone could tell me what to uninstall or remove and then what to install after that would be great.

My full system spec can be found here - http://support.hp.com/us-en/document/c04343334

This is my first time upgrading any components in this computer

game4it70

Did you try the "Intel chipset driver" and the "Intel management engine" driver for HP's site?

I have seen people using older motherboards having issues with the 480's but normally a bios update helps.Not much use to you though.

I presume the card you upgraded from was a typo and is actually a R9 270 ?
If so I'd guess they use about the same power so the PSU should be up to the task.

Gpu fully seated and connection from power supply connected fully?

DarraghR

Think i went through all of the HP updates for my PC but will double check the ones you mentioned.

Yeah cant seem to find official BIOS files

I checked the power situation before buying but when this problem occurred I really thought it was related as it only happened when I jacked the graphics up to Ultra but now it happens either way. And yep it was a typo

It looks to be. Just compared both cards and the only major size difference is the length.

Testing my old card now to see if damage was done to something on the mother board

robcon

What psu have you got Darragh? Model number just so we can rule out power draw as the cause.

Also a very quick search shows links between AMD overdrive and AODDriver2.sys error so remove that from the crimson install and see how you get on.

DarraghR

Intel Core i7 4790
CPU speed: up to 3.6 GHz (max turbo boost - 4.0 GHz)
CPU cores: 4
CPU Cache: 8 MB
TDP: 84 W
Socket: LGA 1150

RossieMan

PSU is power supply, not cpu.

DarraghR

746177-001 HP 500W 24Pin Desktop Power Supply Unit - Delta DPS-500AB-15 A

http://www.ebay.ie/itm/746177-001-HP-500W-24Pin-Desktop-Power-Supply-Unit-Delta-DPS-500AB-15-A-/301785167405?hash=item4643cc362d:g:16QAAOSwZQxW2j1K

RossieMan

A Hp power supply might not be able to run that graphics card. It's advertised as 500w, but chances are it's more like 300-350w.

TerrorFirmer

RossieMan wrote: »

A Hp power supply might not be able to run that graphics card. It's advertised as 500w, but chances are it's more like 300-350w.

Most OEM units do deliver on what they promise as they're typically only included in higher end units that are initially sold as 'gaming PCs' versus your average home PC which only come with 200-250w supplies.

The card he replaced was an R9 270 which actually draws similar power to the RX480 - not ruling out the power supply but the whole system under load still shouldn't be drawing more than about 300-350w ballpark and that supply should be up to the job no problem. HP also used these 500w supplies in systems running HD7950's and GTX670 class cards which are actually more power hungry than RX480 or GTX1060 cards.

The other issue is one of three potential problems:

1) RX cards do have issues with some older motherboards, though yours is pretty recent so it should be fine.

2) HP themselves sometimes (not in all models) tend to lock down card support in the bios, therefore any card released after the last bios may work patchily or not work at all. I've encountered this infuriating situation in many older HP machines, but for some reason always the 'home' type - never the Elite range (which tends to be all the gaming models), or the workstations either.

3) The PSU for some reason is having issues supplying enough power on a single 8-pin. You could try a dual 6-pin to single 8-pin adaptor and see if this helps.

In one machine all GTX6 cards inc the GTX690 worked fine but a GTX750Ti and GTX760 would not even boot, and of course there were no new bios' released after about 2013 for the machine in question.

Also, is there an option in your bios to manually select PCI Express revision/compatibility? There should be a list like "Auto", "Gen 1", and "Gen 2" - try manually selecting Gen 2 or Gen 1 at a time and see if either help, this worked for me on an older board with an RX470 by using "Gen 2".

Also where do you live? If you're near one of us we could have it tested and sorted in an hour either way, whatever the problem is.

DarraghR

I ran a windowed DOOM and opened MSI Afterburner and Core Temp. Picture 1 is at Medium settings. Picture 2 is at hight settings and just before it crashed when I picked up a gun off a dead guy at the very start of the game.







I cant see anything in my BIOS to change that setting but maybe I am missing something. Below a few screenshots

Cuddlesworth

Remove the AMD driver in safemode, reboot, install AMD driver.

If it keeps happening, its probable that the card is faulty.

That Bluescreen basically means that the AMD driver was in a loop.

TerrorFirmer

I would also be inclined to think it could be the board and the fact there are no bios updates available for it in almost 3 years. I have had tonnes of problems with RX cards on older boards (mostly older than Haswell admittedly but one or two S1150 boards too). Nvidia GTX10x0 cards don't seem to have this issue, zero problems here running GTX1050/60/70 on a wide range of S1150 and even S1156 boards.

I'm not saying it definitely is, but I'd see if I could test the card out in a newer machine first before dismissing it as faulty. If you say where you are, someone here might help. It would be literally a 10 minute job to establish if the RX480 was faulty in a newer, ideally Skylake machine. The other option I'd recommend would be to return it and get a GTX1060 6GB instead, as these cards don't have the same compatiability issues that affect the RX cards on older chipsets.

DarraghR

I spent a lot of time researching on what card to buy and to run into this problem is quite frustrating. My monitor is freesync so this is why I went with a AMD card. Some people are starting to say that I can forget about running DOOM at Ultra settings yet the game shows 100+ fps at High settings.

In relation to returning the card is that possible as I have broken the seal? Would they not say that I should have been aware of any incompatibility issues.

Its been 10 years or so since I went near PC gaming but I am already reminded why I didnt invest previously. I know when i get it sorted it will be worth it as I have experienced 30 seconds of DOOM and it looked amazing when compared to my old R9 270

Knocking it on the head for the moment but will try some things mentioned that have not already been tried. My whole day yesterday was spent trying to fix this and till the early hours of the morning (3:30am)

wotzgoingon

You still didn't give your location. Maybe someone nearby could test out the graphics card in their system.

Some how I have to agree with Terror and say it may be incompatible for some reason. Once it is tested in someones system and all is fine then we have narrowed down the possibilities of what is wrong.

DarraghR

Oh sorry I am living near South Circular Road so about 15 minute walk into City Center. Have a car so can drive to anywhere in Dublin. Good thinking. At least if it is a faulty card then I can return it. Cheers guys

wotzgoingon

Well you are no where near me or Terror but you are in a location that there is a lot of users from afaik. Hopefully one of them pipe up and offer to test the card for you.

Lu Tze

wotzgoingon wrote: »

Well you are no where near me or Terror but you are in a location that there is a lot of users from afaik. Hopefully one of them pipe up and offer to test the card for you.

Im in Sandyford, but running an even older socket (1155), so if there are issues with older boards I may not be able to confirm if the card is faulty

2ndcoming

DarraghR wrote: »

I spent a lot of time researching on what card to buy and to run into this problem is quite frustrating. My monitor is freesync so this is why I went with a AMD card. Some people are starting to say that I can forget about running DOOM at Ultra settings yet the game shows 100+ fps at High settings.

In relation to returning the card is that possible as I have broken the seal? Would they not say that I should have been aware of any incompatibility issues.

Its been 10 years or so since I went near PC gaming but I am already reminded why I didnt invest previously. I know when i get it sorted it will be worth it as I have experienced 30 seconds of DOOM and it looked amazing when compared to my old R9 270

Knocking it on the head for the moment but will try some things mentioned that have not already been tried. My whole day yesterday was spent trying to fix this and till the early hours of the morning (3:30am)

For what it's worth I get steady 110fps on Ultra in Vulkan mode with an RX480 and an i3-6300, so the card itself is well able for it, unless of course it's faulty.

DarraghR

Is your Power Limit set to +50 in the ReLive control panel?
Is your computer power plan set to 'Performance'?

I changed these settings and still crashing like mad

K.O.Kiki

DarraghR wrote: »

Is your Power Limit set to +50 in the ReLive control panel?
Is your computer power plan set to 'Performance'?

I changed these settings and still crashing like mad

Did you set these back to +0?

DarraghR

Yes why?

Ok so I just uninstalled every single AMD driver and software. Then I ran Dead Space 2 at max settings without drivers installed. It didnt crash at all. Would this indicate that it is software related and nothing to do with a faulty card or under powered power supply?

K.O.Kiki

DarraghR wrote: »

Yes why?

Ok so I just uninstalled every single AMD driver and software. Then I ran Dead Space 2 at max settings without drivers installed. It didnt crash at all. Would this indicate that it is software related and nothing to do with a faulty card or under powered power supply?

I'd say so, try installing earlier drivers (like 1-2 revisions back) and see if it runs fine.
Also make sure to reboot after every driver install

Cuddlesworth

DarraghR wrote: »

Yes why?

Ok so I just uninstalled every single AMD driver and software. Then I ran Dead Space 2 at max settings without drivers installed. It didnt crash at all. Would this indicate that it is software related and nothing to do with a faulty card or under powered power supply?

Its a tough sell, drivers utilise the card properly. There could be a faulty part of your ram on the card that's only hit with drivers, then without.

If you can game fine with the other card, but swap in the 480 and hit Blue screens constantly, then its most likely a failure with the card.

DarraghR

I dont get that blue screen anymore. Just Every game crashes after a few seconds. I ran Star Wars Battlefront in a windowed mode on Ultra settings and it seemed fine for the most part. Once I went to full screen it crashed. Yet Dead Space 2 will not make it past the start screen ...

I have taken 2 screenshots from GPU-Z. Maybe someone can tell me if they can see anything abnormal

DarraghR

So looks like everything is working alright now. Played DOOM on Ultra for about 20 minutes and GTA 5 for a half hour with no crashing.

What I did was:

1. Uninstalled with AMD Clean Up Utility
2. Uninstalled with DDU
3. Uninstalled my Avira Anti Virus
4. Removed protective cover on desktop
5. Installed the latest AMD drivers

Not too sure which one solved the issue but I would guess uninstalling old drivers with AMD Clean Up and DDU?

pcuser

Daragh I just spotted this thread.

I had a very similar problem to you. Except I would have a black screen instead of a blue screen. Turns out it was the MOBO needing replacing.

Hopefully you have the issue solved but if it happens again it might be worth investigating.

DarraghR

Everything in working order the past 4 days. Must have been a driver issue. Im guessing using both AMD Clean Up Utility and DDU cleared any old driver files. Cheers guys

DarraghR

After installing a new SSD drive my problem is back again. Uninstalling all drivers with AMD Clean Up Utility and DDU has not worked ...

wotzgoingon

Not good. Did you do a clean install of windows on the new SSD?

DarraghR

Yes I formatted the SSD and used a usb made with Windows Media Creation Tool to reinstall Windows 10. Had to unplug my old HDD to do this as it would not let me otherwise. Could it have something to do with my old HDD?

Also under Sapphire TRIXX my VRM temperature looks higher than it did before

DarraghR

Anything look abnormal to anyone?

TerrorFirmer

Use MSI Afterburner or AMD Wattman (it's in the software that comes with the drivers) to underclock your card to 1200mhz core and set a target temperature of 50c or manually set fans to high. See how that helps initially to rule out instability or overheating.

edit - or Trixx.

DarraghR

That shouldn't be necessary. Also power consumption would not be the cause as I have the old HDD disconnected. I guarantee it is a driver issue again. Just dont know what to do this time as my previous method of uninstalling drivers with both DDU and AMD Clean Up Utility wont work.

TerrorFirmer

I never said anything about power consumption. Sometimes a card will be unstable if the core speed is too high. It shouldn't be the case with that model, but it does happen on occassion. Some cards also hard-lock up or driver crash when temperatures get to what's usually considered a safe level due to some issue with vrm overheating or something.

Underclocking the core and manually setting fan speed to say 80% at once (even 100% if the card isn't intolerably loud), then running games and seeing what happens, rules out either possability simultaneously if it still crashes. It's not a permanent solution, it's about ruling out what could be wrong.

At that point, if you pass both with the same situation, the issue is either drivers (seems unlikely if you've already tried different sets of drivers with the same issue) or some sort of difficult to diagnose compatiability problem between the card and board. RX cards do have issues with some older boards, I've personally had that problem a lot.

DarraghR

Sorry I just meant that the power is another one to cancel out not that you were saying it could be the power.

I will give the underclocking and fan speed a go once it does not effect my warrenty.

Before I fitted the SSD I had the same problem with my HDD but solved it so it cant be compatibility with the gpu and board.

Everything is backed up so I wonder should I do another wipe and reinstall of Windows 10.

DarraghR

What about all the drivers for my computer. Do I install every single one of them again? There is a NVIDIA driver listed but it was an AMD card in my computer...

TerrorFirmer

You don't need any of those apart from chipset driver, Win10 should automatically install everything you need once you're connected to the internet. In fact, Win10 automatically installs Nvidia/AMD drivers as well now.

Underclocking or setting fan speeds won't void your warranty.

DarraghR

Intel Graphics Driver for Haswell will not install.

Its says "This computer does not meet the minimum requirements for installing the software"

TerrorFirmer

You don't need that driver.

DarraghR

I reduced the GPU Clock and Memory Clock by quite a bit. I ran Battlegrounds and it did not crash. Check out the screen shots I took in game







The GPU climbed pretty rapidly. I guess the problem is solved but I would like to get the most out of my graphics card. It ran fine previously without any under clocking so why not now

TerrorFirmer

You need a more aggressive fan profile. 73c at an underclocked core is very high - what was happening was plain overheating. Possibly your PC was never designed to have a higher end card in it, and doesn't have adequate airflow. This is a common profile in OEM machines.

I would create a custom fan profile, in your pictures it's at a fairly high 73c despite the lower core clock but the fan is only at 27%.

DarraghR

Would increasing my fan speed but leaving everything else at default keep my warranty intact?

My PC is a gaming machine and is not ridiculously old. I still do not understand why it ran fine before without any tinkering

RossieMan

Your warranty is not affected by your fan speeds, no.
I'm assuming this is our 1st pc?

Google is always your friend.

Moon54

Just skimmed the thread, but is it possible the heatsink was removed from the GPU at some time?
Maybe the thermal paste is poorly applied or a issue with a thermal pad on the VRAM???

I'm just guessing but it sounds to me like high VRAM temps are causing issues.
Have you GPU-Z and monitored VRAM temps under load (they are at the bottom of the Sensors tab).

Also, there's a new driver out yesterday from AMD, Version 17.4.1 - http://support.amd.com/en-us/download/desktop?os=Windows+10+-+64

DarraghR

Its only a new card and never touched the heatsink.

All up to date with the newest drivers. Is it possibly that some driver is causing the high VRAM temps?

K.O.Kiki

@DarraghR
Sounds to me like you have a bad card, though it's strange that it was working for the past 3 months.

Drivers wouldn't cause overheating - that's down to bad thermal design on the VRMs.

Moon54

DarraghR wrote: »

Its only a new card and never touched the heatsink.
All up to date with the newest drivers. Is it possibly that some driver is causing the high VRAM temps?

Ahh, GPUZ doesn't show your VRAM temps but GPU temp looks good.
Maybe try HWiNFO - https://www.hwinfo.com/download.php

Found a Reddit post here that suggest RX 480 users should check their VRAM temps - https://www.reddit.com/r/Amd/comments/4v4zz1/have_any_of_you_checked_your_rx_480_vrms_temp/
At least if the VRAM temps look OK, it'll eliminate temperature as being the issue.

Btw, are you using AMD Wattman? Another redditor suggested that it caused crashs.
To remove it they suggested use DDU then reinstall driver with custom settings and only tick display adapter and HDMI audio.
Worth a try.

DarraghR

Moon54 wrote: »

Ahh, GPUZ doesn't show your VRAM temps but GPU temp looks good.
Maybe try HWiNFO - https://www.hwinfo.com/download.php

Found a Reddit post here that suggest RX 480 users should check their VRAM temps - https://www.reddit.com/r/Amd/comments/4v4zz1/have_any_of_you_checked_your_rx_480_vrms_temp/
At least if the VRAM temps look OK, it'll eliminate temperature as being the issue.

Btw, are you using AMD Wattman? Another redditor suggested that it caused crashs.
To remove it they suggested use DDU then reinstall driver with custom settings and only tick display adapter and HDMI audio.
Worth a try.

Wow I actually totally forgot about Wattman causing issues for other people. Cheers this could be the solution. Will have a look when I get home

DarraghR

Here are my HWiNFO results whilst on the desktop



Not installing Wattman didnt work

I can run games if my fans are cranked up to 80% with slight underclocks to the gpu and memory clock. None of this was necessary before

Redfox25

If  your card is only 3 months old, start the RMA process. Who did you buy from?
It would be a shame to let it be for another while and let your warranty run out. If it happens with the replacement card then your issue is bigger.

DarraghR

I got it form Elara.ie

The problem is I have already sold my previous card and I do some freelance work.