Hi,
For the last few days I am not able to play any games. I thought it had to do with the weather (26°C), but while I was playing BF3 my PC randomly shut down. I started it back up and thought it was a power failure. As I joined the BF3 game again, the same happened, my PC died after 5 minutes. This time however, it refused to start up for the next few minutes.
Now I was playing Banished today, and I saw my GPU temp slowly climb up to 81°C. At this temp, the PC shut down. I restarted and checked my temps using Open Hardware Monitor. My CPU temp is roughly 40°C and I doubt this is the problem.
The GPU I have is an XFX Radeon HD7970 3GB Double Dissipation edition, and I do not have GPU Overdrive enabled in AMD overdrive. I noticed, however, that the speed of my fans do not even get to 100% before my PC fails. In the past I have seen temperatures above 80°C and it did not shut down my PC, but BF3 always caused my GPU fans to run at 100%.
I read multiple websites where everyone puts the PSU at fault but I simply don't see how it runs for a couple of seconds and then stops running. My PC worked just fine a few weeks back, I have been away for 12 days and when I came back I only did some programming, no gaming. The situation or place my PC is located has not changed so I doubt this is the cause. I have a case with 4 fans blowing air in and 2 fans blowing air out. My CPU (AMD FX8150 @ 4.1 GHz) has a Scythe Mugen 3 cooler. I checked all fans (CPU, GPU, PSU) and they are all spinning. The GPU and PSU are barely a year old so I guess it is covered by warranty in case it is at fault.
I have an XFX Core Edition PRO750W power supply, I guess that is important information aswell.
Wow, 81 C, my Gigabyte Radeon HD7870 and my Evga GTX 760 have never gotten that hot. High 60s at the most when maxing out BF4 (with my GTX 760).
This is not the solution to your problem, but maybe you should check your GPU thermal paste while you're troubleshooting, or maybe you just need to manual adjust your fans with the AMD Catalyst.
[QUOTE=DEG_fan;45500720]Wow, 81 C, my Gigabyte Radeon HD7870 and my Evga GTX 760 have never gotten that hot. High 60s at the most when maxing out BF4 (with my GTX 760).
This is not the solution to your problem, but maybe you should check your GPU thermal paste while you're troubleshooting?[/QUOTE]
I'm not sure how I would do that. Like I said, the GPU is not even a year old and I assume it still has warranty so I don't want to take it apart. My GPU always got that hot and I had the same problem with my old GPU (HD6870).
I can show some pictures of my case if that'd help? Just a quick note, today has been quite "cold" and yet it overheats. I installed new drives between it not overheating and it overheating, but I just installed the new beta drivers and it didn't really solve a thing. I'm not sure if drives could be the cause of this.
Oh and also useful to know, I clean out my PC regularly with compressed air, so I am sure there is not a lot of dust between my cooling block / fans.
-edit-
I have just been playing Garry's Mod for a good 2 or 3 hours. Here's what my Open Hardware Monitor captured:
[img]http://ss.rubensrv.nl/i/1264gV.png[/img]
80c on the gpu is quite high. I'm pretty sure the 7970 cards start to become unstable at 74c
Try to get some better cooling on the cards, add more case fans. Manually set the gpu fans to 90 or 100% you need to get your max temp on those cards to be 70c under full load with furmark
If it is the card gigabyte is really good with rma's I am a 7970 / 280x murderer (killed 3 or 4 cards) they've replaced or repaired my cards every time and quickly.
[QUOTE=Blackha7;45504612]80c on the gpu is quite high. I'm pretty sure the 7970 cards start to become unstable at 74c
Try to get some better cooling on the cards, add more case fans. Manually set the gpu fans to 90 or 100% you need to get your max temp on those cards to be 70c under full load with furmark
If it is the card gigabyte is really good with rma's I am a 7970 / 280x murderer (killed 3 or 4 cards) they've replaced or repaired my cards every time and quickly.[/QUOTE]
If I manually set the fans to 100% before I start playing and keep them that way, it still overheats and then the PC cuts power off. So even 100% fans won't help me. Either way, I have 6 case fans, 4 going in, 2 going out, and there's an air cooler right next to my PC so there is enough cold air. I guess I need an RMA?
[editline]26th July 2014[/editline]
[t]http://ss.rubensrv.nl/i/1264ZB.jpg[/t]
The fans on the side blow in, the one on top blows out
[t]http://ss.rubensrv.nl/i/1265X4.jpg[/t]
The fan on the right bottom (blue light) is also blowing air in, fan on the left top (back) blows out
[t]http://ss.rubensrv.nl/i/1266Pk.jpg[/t]
The GPU fans are not blocked and should have an almost clear way to get air out.
There is no other place for all the wires on the bottom, so I can't do anything about that.
-edit-
I had to cut off the FutureMark demo. My temps were raising too high, this is a screenshot 2-3 seconds after shutting down FutureMark and after FutureMark had run for not even 20 seconds:
[t]http://ss.rubensrv.nl/i/1267W4.png[/t]
Worth mentioning is that you can see that the fans didn't even reach 100% yet, even though the temp is > 70
I just ran FutureMark again, and after 58 seconds my GPU reached 81 my PC's power cut off. Fan was manually set to 100%. Idle temp was 53.
You will probably need to rma the card. Tell them it is overheating.
I have heard xfx sometimes uses too much thermal paste.
Could possibly be the psu but with a hot gpu there is no way to accurately test the psu.
With that much cooling/air flow you have going on there your GPU should not being getting that hot. I read quite a few reviews where people complained about that particular card running a bit too hot in games that shouldn't strain it so.
I emailed the company I always order my hardware. I explained them my situation and I hope I'll get an email back by monday. If they tell me I should RMA, I will RMA. I first came to Facepunch though, bit worthless sending my GPU on RMA if it is a trivial fix.
Thanks for the help though.
if it's serious enough you could also underclock your GPU for a while until you can fix it
it'll make things slower but you won't deal with heat, just remember to be careful not to go crazy
not a solution to it at all though, but at least you can play games if you did.
[t]http://ss.rubensrv.nl/i/1264vC.jpg[/t]
Call me ridiculous (it probably is) but I put an aircooler (not an airconditioner) next to my PC. Instead of idling at 52-53 degrees, it now idled at 42 for a little. It definitely cools it down, but now my PC shut down while my GPU was at 75 degrees already. The GPU load was at 98%. How can I check if it is indeed not my PSU? I do think my GPU got warmer than 81 before, but I can not confirm.
Maybe you should try doing the paperclip test on your PSU. You can check to see if your fan is still spinning after a certain amount of time. Maybe the fan is not spinning and causing your PC to crash/overheat.
[QUOTE=DEG_fan;45508279]Maybe you should try doing the paperclip test on your PSU. You can check to see if your fan is still spinning after a certain amount of time. Maybe the fan is not spinning and causing your PC to crash/overheat.[/QUOTE]
Any videos?
-edit-
Sorry, now I have a bit more time, I reread your post and I can assure you that the paperclip test won't help me (I already looked up a few videos on Youtube). The thing is, I am 100% certain my fan is spinning. I can keep my PC up and running, playing Garry's Mod and such for 6 hours straight. It really happens when my GPU load is 98% and stays like that for a while. The temperature might be just a coincidence the first few times, considering that now I had my PC crash at temperatures around 70 C.
I'm starting to doubt if it's my PSU or my GPU. A friend of me has his PC with him (he's currently here). I might switch my GPU with his and see which is the culprit. He's running two 6870's (one of them is my old one). He has a 750Watt PSU just like me, so I hope it will find out what the cause is.
-edit2-
I just tried my friend's card. Also an XFX one, but 6870. The recommended wattage for the 6870 is 550w, and the wattage for the 7970 is 650w. So we're not sure if it is a wattage problem. Either way, we are now going to test my card in my friends PC and his power supply, also 750w.`
-edit3-
Just tried my own card in my friend's PC. Same behavior, but for some reason his benchmark runs way smoother. He's only got a different motherboard (ASRock 970 Extreme4 vs ASUS M5A78L-M/USB3) and his CPU (AMD FX-8130) with a lower clock runs the benchmark smoother with my card than my card does in my PC.
80C isn't the end of a world for a GPU, nor a CPU. I just think the card itself might be having issues when under pressure, you could try underclocking it?
[QUOTE=Killervalon;45518281]80C isn't the end of a world for a GPU, nor a CPU. I just think the card itself might be having issues when under pressure, you could try underclocking it?[/QUOTE]
I already underclocked it to 825 MHz. Didn't do any good. However, I don't think it should be necessary to underclock a GPU when it is running at 925 MHz at stock. As far as I know, there are other vendors (So, not XFX but others) that run the same chip at 1025 MHz, without problems. So I already have a lower clocked one by default, it gets > 80 C and it fails under load.
I just RMA'd my GPU today. Hopefully will get an answer soon. What happens if my card is broken? Do they send it to XFX or will I just get a replacement?
(I mean, what is the usual case, I know you can't know it for sure)
They acknowledged my problem and now are sending my card to XFX. Can take up to 4 weeks :(
Would it be possible that your cpu heatsink is displacing heat directly onto your card. Also I recommend installing speed fan and checking to see if all of your fans are spinning and cool etc.
I see you have your card on the way but a little experimenting wouldn't help until then.
[QUOTE=Hollosoulja;45644375]Would it be possible that your cpu heatsink is displacing heat directly onto your card. Also I recommend installing speed fan and checking to see if all of your fans are spinning and cool etc.
I see you have your card on the way but a little experimenting wouldn't help until then.[/QUOTE]
CPU heatsink blows up, directly out of the case.
I just got an email saying they will credit my GPU. XFX can't repair it I suppose. They offered me the "MSI R9 280X GAMING 3G" as a replacement. Should I accept this?
[QUOTE=Cyberuben;45696767]I just got an email saying they will credit my GPU. XFX can't repair it I suppose. They offered me the "MSI R9 280X GAMING 3G" as a replacement. Should I accept this?[/QUOTE]
Yes
[QUOTE=Killervalon;45697653]Yes[/QUOTE]
Thanks. I just accepted it, but about 5 minutes ago they mailed me saying that when looking at the specs, the MSI card has a 75 MHz higher clockspeed.
[QUOTE=Cyberuben;45698010]Thanks. I just accepted it, but about 5 minutes ago they mailed me saying that when looking at the specs, the MSI card has a 75 MHz higher clockspeed.[/QUOTE]
75mhz is just a move on a slider and it's done. The difference is minimal and you can do it yourself.
[QUOTE=Killervalon;45710177]75mhz is just a move on a slider and it's done. The difference is minimal and you can do it yourself.[/QUOTE]
I never felt safe OCing my GPU because of overheating, so even if it's just a slider, if the manufactorer did it I know I will be fine.
Okay so I just received my GPU yesterday. I can play at 100% load just fine, GPU temps stay below 65 degrees (which is just a bit higher than my old card's idle temperature). Battlefield 3 runs at 70-120 FPS and Battlefield 4 runs at 55-110 FPS on custom highest settings.
Thanks for the help everyone, thread resolved!
Sorry, you need to Log In to post a reply to this thread.