• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

R9 380 high temperature and GPU core at idle

Flow said:
Run gpuz and check gpu load in second tab, should be around zero when idle. And don't mind the jumping, it will do that because of anything you have open on your desktop.
Basically it should report zero every other moment.
You can use cpuz and realtemp to check cpu speed and temps, are those idling correctly?
If you open the ccc drivers and go to performance/amd overdrive you can enable that and leave everything at default. Maybe your card doesn't idle correctly so worth a try.
Does your card reach the correct speeds when running a game or benchmark?
Does the card behave the same in windows safe mode?

So if everything says your at zero activity for the gpu, it could be your cards bios is faulty or configured that way since it's in that where the powerstates are defined.

On a side note, some animation or video could be causing this, even when closed, maybe something that starts with windows and keeping the gpu in the wrong state?
Click to expand...

Thanks Flow. CPU speed and temps are idling correctly. I'll try your suggestions tonight!
 
Improve airflow to or over the GPU in the case.
If the card is at idle and that warm it could be lack of flow keeping it warm.
 
65C is not hot for a graphics card. My PowerColor R7 360 does the same. PowerColor cards seem to have the fanspeed target 65C...you should never see the card go much over that.

Some cards are incapable of reducing their memory clock at idle...some are. Both of my cards are either at full clocks or 300/150.

390 is 42C at idle (20% fan speed)
360 is 30C at idle

Look in Catalyst for current GPU info. Open the Crimson panel, Gaming, Global Settings, then Global Overdrive [your card]. It takes very little activity to get the memory clocks up. GPU should be mostly at 300 MHz if nothing is truely going on.


Caring1 said:
Improve airflow to or over the GPU in the case.
If the card is at idle and that warm it could be lack of flow keeping it warm.
Click to expand...
Yup, I'd verify the fans are spinning too--they should be or it would exceed 65C.

I'm thinking 0% activity + 65C = HSF improperly mounted or something wrong with the thermal transfer material. If this is the case, return it to where you bought it for a replacement.

I've had lots of PowerColor cards and they're all excellent in terms of thermals.
 
Last edited:
Thanks everyone! Unfortunately the card is still doing the same thing. :(

I uninstalled my drivers and then I ran DDU. I also removed anything and everything AMD/ATI.
and then I installed the drivers.
I tried this process with both the drivers on the CD that came with my graphics card as well as the latest drivers on the AMD website.


I downloaded Process Explorer and all I see using the GPU is explorer, Shell Experience Host and Cortana. I don't think they're putting much load on it though.

ProcessExplorer.jpg



I also checked GPU-Z and the GPU load stayed mostly at 0% when I'm not doing anything.

GPU-Z.jpg
 
well ... what about repasting it?
 
That wont lower the clocks, which seem stuck at those speeds when idle.

http://www.bjorn3d.com/2015/07/powe...r9-380-4gbd5-ppdhe-affordable-4gb-solution/3/

Reading the review I notice your clocks should be higher at full load, but lower at idle, your temps are as if it's under load.
Try to see if it's also that high when in safe mode, only other option would be a clean reinstall of windows and then checking the gpu core and memory speeds.

2.gif
3.gif

1.gif


One is with a youtube video running, only memory up at 1600Mhz.
One is full idle and one is full load from furmark for a quick reference.
Btw, my real idle temp for gpu is around 30, webbrowsing and the likes do make it a bit warmer which shows in the idle pic.
 
Last edited:
Subbing to see how this turns out. Welcome to the forums, keyjey! Hope this gets fixed.

Out of curiosity, is there a second PC you could put this in to see if it experiences the same issue?
Also, have you tried putting the GPU under load to confirm the clocks move up at least?
 
A flash of the bios shouls fix this.
I have my Vapor 280x sitting at 65C at load, 50% fans.
My card has 3 modes in the factory bios:
1. Idle: 150 mhz / 300 mhz
2. Idk why this profile is here: 700 mhz / 1500 mhz
3. Full clocks: 1070 mhz / 1500 mhz.
I guess your gpu got stuck in the second mode like i have on my 280x and it runs on it all the time.
See in drivers if power save is disabled or something like that.
 
Filip Georgievski said:
A flash of the bios shouls fix this.
I have my Vapor 280x sitting at 65C at load, 50% fans.
My card has 3 modes in the factory bios:
1. Idle: 150 mhz / 300 mhz
2. Idk why this profile is here: 700 mhz / 1500 mhz
3. Full clocks: 1070 mhz / 1500 mhz.
I guess your gpu got stuck in the second mode like i have on my 280x and it runs on it all the time.
See in drivers if power save is disabled or something like that.
Click to expand...

That may be the problem but if it was my card, I would not flash it myself, I would just RMA it.
 
You can always try hitting RESET+Yes in the Overdrive. If the clocks don't move and you eliminated external programs. I would just RMA it.
 
xvi said:
Subbing to see how this turns out. Welcome to the forums, keyjey! Hope this gets fixed.

Out of curiosity, is there a second PC you could put this in to see if it experiences the same issue?
Also, have you tried putting the GPU under load to confirm the clocks move up at least?
Click to expand...

Thanks xvi. I do have another computer but the power supply on that does not work with this gpu. :( And moving the power supply is not an option since the motherboard uses a different connector.

I ran FurMark as well as a few games and the Core clock would move up to 980MHz. Memory clock stayed the same though.

Filip Georgievski said:
A flash of the bios shouls fix this.
I have my Vapor 280x sitting at 65C at load, 50% fans.
My card has 3 modes in the factory bios:
1. Idle: 150 mhz / 300 mhz
2. Idk why this profile is here: 700 mhz / 1500 mhz
3. Full clocks: 1070 mhz / 1500 mhz.
I guess your gpu got stuck in the second mode like i have on my 280x and it runs on it all the time.
See in drivers if power save is disabled or something like that.
Click to expand...

Thanks Filip! That's really interesting. But I'm not comfortable flashing the bios on my GPU.

Thanks P4-630 and Xzibit. I'm requesting RMA now. I'm doubtful but hopefully it's a quick process :D
 
Last edited by a moderator:
  • Like
Reactions: xvi
cdawall said:
Is ulps turned off? It should drop to the 1xx-3xx range for single 1080p 2d
Click to expand...
Stop making this recommendation already! We've have enough discussions about it where this shouldn't even be coming up. ULPS is basically a sub-idle mode on secondary CFX cards for people who run CFX . If you only have one GPU, your GPU will never enter ULPS because it wouldn't be able to drive anything. ULPS only impacts secondary CFX cards by practically turning off everything.

@keyjey: Please stop replying within minutes of each reply. You should be using the "edit" button if no one has yet responded to you. Moderators tend to get frustrated with posting multiple times like in the matter you just did. Just update your last post if no one has replied yet.
 
Mode 2 is for playing video files. My 7970 would do 500Mhz on the gpu when playing a video. My current card only sets memory higher when viewing a video.
Also, if the 380 is as my 390, it wont have a bios switch, making it a serious risk to flash it with something that possibly wont work and could brick the card. Rma is the best answer and solution.
Goodluck and hopefully you get it back quick.
Also fingers crossed it's the card indeed, and not some registry setting faulup from windows.
 
keyjey said:
Hi everyone,

I just bought a Powercolor PCS+ r9 380 4GB graphics card and I noticed that on idle my GPU temperature is at 64-65C and GPU core is at 755MHz. The fan runs at around 39%. The GPU core never goes below that. I'm running a single 1080p 60Hz monitor.

Any idea how I can lower the GPU core at idle?

Thanks!
Click to expand...
MSI Gaming APP it is made for this GPU ..
Asus rock might have amd mobo drivers to help with the amd video card ,I had to up date all drivers .from MSI to use my R9 380 8 GB now i see AMD ACP Binaries ,and a few things new . And i always use amd drivers ... http://support.amd.com/en-us/kb-art...mson-Edition-16.1.1-Hotfix-Release-Notes.aspx
with the hot fix for over heating ...
[/FONT]
 
Flow said:
That wont lower the clocks, which seem stuck at those speeds when idle.

http://www.bjorn3d.com/2015/07/powe...r9-380-4gbd5-ppdhe-affordable-4gb-solution/3/

Reading the review I notice your clocks should be higher at full load, but lower at idle, your temps are as if it's under load.
Try to see if it's also that high when in safe mode, only other option would be a clean reinstall of windows and then checking the gpu core and memory speeds.

View attachment 72019View attachment 72020
View attachment 72018

One is with a youtube video running, only memory up at 1600Mhz.
One is full idle and one is full load from furmark for a quick reference.
Btw, my real idle temp for gpu is around 30, web browsing and the likes do make it a bit warmer which shows in the idle pic.
Click to expand...
Any chance you have use overclocking program's ,if so i fond out that having more then 1 it will lead the motherboard and os mes up and not regulate power correctly i would flash you bois ,it could still have old data from your old video card and all drivers need to be uninstalled '' if you did not change os when you got video card '' and look for any data in your User AppData and msconfig to see if things are still starting that you uninstalled . amd overdrive and is a good one for that ,good luck you have a great video card hop you get it to work.
 
You must log in or register to reply here.
Back
Top