Two 4870X2 Crashing and "Boiling" !!

Di7

Distinguished
Apr 21, 2009
77
0
18,630
Hello..

Yesterday, I bought two 4870X2 Cards and when I installed them the Operating System crashed and rebooted.

I'm using Windows Xp x64 and Windows Vista Ultimate x64 and both systems are facing the same problem.. I tried to use the latest drivers versions and Catalyst CC 9.4 (I also tried to install the drivers only using windows).

When I check the temperature I realize that it is 81-86C and up for the First GPU.. the other GPUs are no batter, about 74-81C..!!

I tried to disable the CrossfireX.. use older drivers.. but no improvement..

At the End I decided to remove one of the cards.. the OS takes longer time to crash but I realize that the temperature is the same (80C and above) even that I'm using a single card.. Even that the activity is ZERO.

By the way, I disconnected most of my hard disks and devices to avoid any interruption. ( Just the RAM the CPU the GPU and a single HDD are ON).

Look at these Pictures:












Computer specifications:
CPU: Intel i7 Extreme 965 3.2Ghz.
Motherboard: Gigabyte EX58-Extreme rev.1
GPU: TWO ATi RADEON HD 4870X2 2GB
Memory: OCZ 1600 (6X2GB) 12GB Total.
HDD: 3 WesternDigial HDs with 2TB (6TB Total using RAID). and 2 WD 500GB SATA2 (inside, in the floppy place). 2 other SATA2 HDs outside ( 7 HDDs total).
Blu-ray Burner: LG BRD Burner
Case: Gigabyte-Poseidon-310
Power Supply 1: COOLMAX GreenPower 950 watt. (is it enough or should I use the 2nd PS all the time?)
Power Supply 2: COOLMAX GreenPower 1200 watt. (for Ultimate use: when All HDDs and devices are on)
 

Euphoria_MK

Distinguished
Mar 4, 2009
492
0
18,790
Hmm dont know what to tell you... Poor air flow in your case?

I just received my 2nd 4870 1GB Vapor-X card today and I already tested it and I got 18960 points in 3DMark 2k6 and played Left4Dead all maxed out without any problems.

I know that the 4870x2 and 4850x2 cards create more heat than the single GPU cards, but they should be fine.... Personally I haven't tested QuadFire, but it sounds like you have an awesome setup as long as you resolve this issue.


I have Corsair 750W PSU and Phenom II 940 @3.515GHz, G.Skill 4GB 2x2 @ 1066

 

cokenbeer

Distinguished
Jun 28, 2008
201
0
18,680
Those cards do run hot, but I would also suspect that you have poor airflow in your case if they are running that hot at idle. I had a similar setup and ended up selling the 4870x2 because my system kept crashing. The person who purchased them had no problems whatsoever. Seems to me that depending on your setup there may be some conflicts which cause problems.

Also, if you had an Nvidia card before make sure you don't have the drivers still installed.
 

Di7

Distinguished
Apr 21, 2009
77
0
18,630


Hey..

At the beginning I installed a single card and I tried to play GTA IV with the highest options available and the game was great but I didn't check the temperature at that time.

I tried PES2009, Crysis , and many other games with no crashes.. (maybe the temperature was high but I didn't notice).

The Case was closed when I ran GTA IV at it highest level but now I'm leaving it opened with only one HD inside.

The fans of this case are working just fine it's Gigabyte-Poseidon-310. I know had to buy a bigger case but this one is giving an acceptable space and cooling conditions.



 

jassao

Distinguished
Mar 18, 2009
103
0
18,680
I had temps like that with my ati 4850x2 before, idle at 75-80C load 95C.

What worked for me was to update everything, and in the proper order.
1. Check bios for correct frecuencies and voltages. be sure your pci-e has 100mhz frecuency (it should be the default) and pci's at 33.6mhz. make it all default.

2. Clean install vista x64 and before any automatic update from vista, update your mobo drivers (download the newest ones for your pc). restart.

3. Update vista, fully. Be sure it installed all DirectX drivers, framework, etc. So you're now under Vista 64bits SP1. restart as many times as you need.

4. Install ATI drivers, Be sure you have your PCI-E slots( for the 2 slots) drivers installed, Just before this step you can put the 2nd card on the mobo and vista will tell you new hardware found, updating pci-e drivers.

If at this point you still see high temps, try updating your bios.

now my card idles at 44C and 75C under load.

And maybe, as said before, check first your case airflow, if you have your HDD's too hot that's a sign that the same hot air is revolving in your case.
 

Di7

Distinguished
Apr 21, 2009
77
0
18,630


Sweetheart... I'm using EXTREME motherboard of Gigabyte (EX58-EXTREME) and it is naked right now with no case just the CPU the RAM the GPUs and a single HD..


I'm running my Computer in front of the conditional.. :) I'll bring it a cold juice too..
 

Di7

Distinguished
Apr 21, 2009
77
0
18,630



Thanks for sharing your experience..

I'll check the bios and see..
 

cokenbeer

Distinguished
Jun 28, 2008
201
0
18,680


Yes, because that "EXTREME" motherboard DEFINITELY means there couldn't be any problems at all. I had a Rampage Extreme, and a P5E3 Premium before and it still had similar problems.

Don't come asking for advice if you don't want any. Try a fresh install of everything, no overclock, update motherboard bios, and see is that works.
 

Di7

Distinguished
Apr 21, 2009
77
0
18,630
I have the latest Motherboard bios F6..

That's bad.. it crashes now during boot and with no activities..

I don't want to replace the cards. Do you think I should?
 

cokenbeer

Distinguished
Jun 28, 2008
201
0
18,680


Try swapping for the other card and see if you still have problems. I ended up going to 4870's in crossfire... less performance but it's stable. Then I went to 285's in SLI, no problems at all and it's ridiculously fast even at 2560 x 1600. I still use ATI, and I would try an x2 card again in a heartbeat but I had a poor experience with the 4870x2. I should note that I gave one card away and sold the other. Both are still working without any issues. Go figure.
 

jassao

Distinguished
Mar 18, 2009
103
0
18,680
If you have the lastest driver/bios for your mobo then I just see 2 possible solutions.

- the hardware is fine and you can try a clean install, use one of your disks and try a new vista x64 installation following the order i mentioned before (even uninstalling/installing the ATI drivers would not work, since sometimes they don't uninstall completely and then they create problems, like yours).

- Software-wise everything is ok, but even if you see fan spinning at 30% on your screen, the fans are actually spinning really slow (sometimes they even stop completely and you see them at 90% and 80C!), or the heatsink is not doing proper contact with the GPU.

Also, is this happening with both 4870x2 or just the one on the screenshot? If this is happening in both cards, most likely is a driver/software problem.
 

Upid

Distinguished
Apr 20, 2009
16
0
18,510
Like jassao said, your fan speed is too low . Enable manual fan control and turn the fan speed to hundred percent for all the gpus . And in crossfirex panel how many gpus have you enabled to crossfirex ?you have to set four gpus there.
 
100%?

Heck no. I have a pair of 4870x2s - those temps are normal, and if you're really concerned, turn the fans to 50%. 4870x2s have vacuum cleaners for fans - if you turn them to 100%, you will need earplugs.
 

Upid

Distinguished
Apr 20, 2009
16
0
18,510
I am saying this only to find out if temp is the problem.otherwise setting speed to 60 or 70 isn't that bothering.
 

Di7

Distinguished
Apr 21, 2009
77
0
18,630


I tried to install windows Xp x64 again but in a usb HD and I faced the blue screen.

The Pictures above when 4 GPUs are ON with crossfire.





 

Di7

Distinguished
Apr 21, 2009
77
0
18,630


81C and you say the temps are NORMAL!!

The temps may go above that.. I'm not sure but the OS crashes due the heat and reboot.
 

Di7

Distinguished
Apr 21, 2009
77
0
18,630


I installed a single card and I turned the fan to 100% (I can't stand the sound of the fans, it's like a plane!!) and the main GPU temperature go down to 34% but when I disabled manual fan control temps reached 68% then gradually to 81% 86% then it crashed (in few hours). (this with a single 4870X2 card).
 

Di7

Distinguished
Apr 21, 2009
77
0
18,630

Look at the top.. the computer specifications.
I'm using 1200 watt. the same performance with 950 watt so I'm stuck with the second one now.
I think 950 watt is more than enough cause I'm not using all my devices at the moment.

Power Supply 1: COOLMAX GreenPower 950 watt.
Power Supply 2: COOLMAX GreenPower 1200 watt.
Total: 2150 watt. (I'm not using them both even that I can, no use).
 

Di7

Distinguished
Apr 21, 2009
77
0
18,630


Nope. SATA cables are far away from the cards.. The nearest cables are the cards' power cables, of course, they are plugged at the top of them.

No I don't think that they effect anything.

I tried to run my motherboard outside the case with the two GPU cards CPU RAM HDD but the same temps..

I'm using Poseidon 310 Case:

http://www.gigabyte.com.tw/FileList/WebPage/thermal_file/poseidon_flash.swf
http://www.gigabyte.com.tw/Products/Chassis/Products_Spec.aspx?ProductID=2677

1625og4.jpg


152i6c7.jpg


flf3gi.jpg


I know it is small for what I have but I don't think that it's the cause of temps problem.

Should I use extra liquid/air cooling systems?
 

Di7

Distinguished
Apr 21, 2009
77
0
18,630
I received support from AMD:

Dear Customer,

Your service request : SR #{ticketno:[8200191761]} has been reviewed and updated.

Response and Service Request History:

Hello,

I understand you system is crashing or receiving irregular shapes/textures.
Lets try the following:

--=-- replace the card with another if possible - does the issue follow

--=-- try the card in another system if possible - does the issue follow

--=-- uninstall the drivers and update the drivers (http://www.ati.com/drivers)
Steps for a manual uninstall are at the bottom of this guide
Uninstall Steps = http://support.amd.com/us/kbarticles/Pages/737-20561RemovingOldATISoftware.aspx
Installation Steps = http://support.amd.com/us/kbarticles/Pages/737-20870HowtoInstallYourATIProduct.aspx
Installation Steps (Vista) = http://support.amd.com/us/kbarticles/Pages/GPU-1-ATICatalystSuiteDriverWV.aspx

--=-- update motherboard chipset drivers
http://www.playtool.com/pages/chipsetdrivers/chipset.html

--=-- remove any USB devices

--=-- try a lower resolution

--=-- remove second display if you have one attached

--=-- try an older version of Catalyst Suite from the older drivers (try 8.2)

If the issue is resolved with a driver change, this sounds like an issue that needs to be addressed in future releases of our drivers and software. We have a form designed to assist troubleshooting and reproducing issues like this to resolve in updates. The form is located at http://support. amd.com/us/kbarticles/Pages/CatalystCrewFeedback.aspx For now, since we know the limitations of the problem, lets stay with the best workaround for now. If or when the issue will be resolved is unknown. Sorry for any inconveniences this may cause.

--=-- visually inspect the fan for speed and obstuctions

--=-- use MSCONFIG to run a clean boot

MSCONFIG is a tool we can use to target is a service or application is causing the problem. The idea is to remove third party items from start-up. If the issue doesn't happen we know one of those items removed was causing the problem. We will then systematically return some items until the issue returns.

1) From the Windows Taskbar, select "START" then "RUN" (Vista removed "RUN" by default, but Start Search will work as well. To get "RUN" to appear, right click on the Start button and select properties > Start Menu > Customize > Select [Run Command] > [OK] )
2) The Run dialog box will then open.
3) In the Open field type "MSCONFIG" and then select "OK".
4) The System Configuration Utility dialog box will launch.
5) Select the "Startup" tab.
6) The startup tab will list all applications that load on system startup.
7) Make a note of all the startup items that are checked. Once you documented all startup items, select the "Disable All" button.
8) Select the "Services" tab.
9) Place a check in the "Hide All Microsoft Services" box.
10) This will hide all Microsoft services and only list third party manufacturer services.
11) Make a note of all the non-Microsoft services that are checked.
12) Select the "Disable All" button. Doing this will uncheck all third party services.
13) Select "Apply" and then "OK".
14) Restart once prompted.

--=-- check if your system has a power supply powerful enough to run the device

--=-- go through our steps for general resolution
http:// support.amd.com/us/kbarticles/Pages/TroubleshootSlowPerformance.aspx

--=-- download MPS Reports Setup/Perf and run them:
XP = http://www.microsoft.com/downloads/details.aspx? FamilyID=CEBF3C7C-7CA5-408F-88B7-F9C79B7306C0&displaylang=en
1) Attach the CAB file to this ticket that it generates
2) the Cab file will be located C:\Windows\MPSReports\Setup\Reports\cab

Vista and run as administrator:
--=-- http://www.microsoft.com/downloads/details.aspx?familyid=00AD0EAC- 720F-4441-9EF6-EA9F657B5C2F&displaylang=en
1) Attach the CAB file to this ticket that it generates
2) the Cab file will be located C:\Windows\MPSReports\Setup\Reports\cab

--=-- Create a problem report:
http://support.amd.com/us/kbarticles/Pages/737-1696HowtoCreateaProblemReportinWindows.aspx

If the above steps show the card is at fault, submit a warranty replacement using http://support.amd.com/us/kbarticles/Pages/737-19285Warranty ServiceRequest.aspx (must be a Built-By ATI device)
--=-- Partner devices go http://support.amd.com/us/partner/Pages/ SupportWarrantyServices.aspx

--=-- Everest Report
1) download - http://www.softpedia.com/progDownload/Everest-Home-Edition-Download-16369.html
2) extract files and run application - everest.exe
3) select report > next > all pages > next > html > next >
4) save to file and attach document to this ticket

Manual uninstall drivers and software

--= As a reference a GUID is {055EE59D-217B-43A7-ABFF-507B966405D8} ==-

Try the following steps to remove and reinstall the software and drivers
1) Remove all ATI software from Programs and Features section (i.e. Catalyst Install Manager).

This includes:
- ATI Software Uninstall Utility
- ATI Catalyst Control Center
- ATI Decoder
- ATI Display Adapter
- Catalyst Media Center
- and so on

2) Go to registry and remove any ATI entries left under:

HKEY_LOCAL_MACHINE\Software and HKEY_CURRENT_USER\Software

HKEY_LOCAL_MACHINE\Software\Microsoft\Windows\CurrentVersion\Uninstall and look at each GUIDs display name for ATI entries remove the GUID if found. (NOTE - click on the first GUID/Long_Number in the list and use the arrow key on the keyboard to go to the next. Look at the right side DisplayName = ATI or something simular)

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Installer\UserData\S-1-5-18\Products look at each GUID and the InstallProperties key, look at the Display Name for any ATI entry, if found remove the whole GUID.

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Installer\UserData\S-1-5-21-61356107-1988573696-2098142674-6793\Products look at each GUID and the InstallProperties key, look at the Display Name for any ATI entry, if found remove the whole GUID

3) Exit Registry editor.

4) Go to Computer and select Organize and choose File and Folder options, Click the View tab. Make sure view hidden files and folder is checked.

5) Now browse to C:\Windows\INF and remove any OEM#.inf (ie OEM0.inf, OEM1.inf, OEM3.inf) and the corresponding OEM#.pnf files that reference ATI. You look only in the OEM#.inf file. The first few lines in the OEM#.inf file will say ATI if it is related to ATI otherwise, leave the file and prceed to the next one.

6) Next remove the C:\Program Files\ATI Multimedia Center folder and C:\Program Files\ATI Technologies folders if present.

7) Restart

8) Update Windows Installer 4.5:
http://www.microsoft.com/downloads/details.aspx?FamilyId=5A58B56F-60B6-4412-95B9-54D056D6F9F4&displaylang=en

When you are reinstalling the software and drivers, stop and disable any anti virus or simular application

9) Reinstall Catalyst Control Suite

NOT SUPPORTED STEP -- The changes are not supported by ATI and may cause adverse
effects to your system if you are unaware certain registry skills
--=-- http://www.microsoft.com/whdc/device/display/wddm_timeout.mspx

In order to update this service request, please respond, leaving the service request reference intact.

Best regards,

AMD Global Customer Care

Older CCC 8.2 driver version has 32bit Edition no 64bit..

http://www.filehippo.com/download_ati_catalyst_xp_64/


 

jassao

Distinguished
Mar 18, 2009
103
0
18,680


What I meant was that sometimes the screen could be saying "fan @30%" and then when you check the fan itself, it's actually spinning really slow, you could move the fan control to 100% and the fan is spinning at the same slow speed.

Since this is not happening to you (because you hear the airplane sound @ 100% fan speed lol) the fans are doing their job, so it's either a driver problem or a Heatsink problem.

I would recommend (instead of all the uninstall/install/testing AMD is suggesting) to do a clean install on a HDD, don't know why you received blue screen from the usb HDD, and try CCC 9.2 64bits, that one worked for me.

I'm sorry, I was assuming you were using vista x64, and I just read you tried XP 64x. Maybe that's the problem with your drivers. I havent tried XP 64x on my system but I've heard horror stories from that OS. My opinion is that a person should use (if you have the option to choose) either XP 32bits or Vista 64bits. Thats just my personal opinion and I understand sometimes you're stuck with an OS.

I think you have vista 64x dont you? so try a clean install of it and see what happens.

Good luck!

Edit: IF it does lower your temp to 34C at 100% fan speed then the problem most likely is with software, the fans are not kicking correctly. That means the software is not sending correct info to the fans, something like "CPU: Oh the GPU is at 80C! -- CCC: Don't worry, keep the fan at 3% we should be fine"
 


Clearly you don't own a 4870x2. Anything above 40% is starting to get noticeably loud, and 60-70% is quite obnoxious. It's not like the fan on Nvidia cards or anything like that. Of course, it also moves a phenomenal quantity of air at high speeds too - it isn't just noisy and useless.



Yes - easily. My GPUs idle at 70+C, and hit 90C in gaming. It's nothing unheard of at all. 4870x2s run quite hot, and it shouldn't harm them in any way.
 

AgoniesFury

Distinguished
Nov 28, 2008
3
0
18,510
di7 i understand your pain because on my prior build i ran pair of 4870x2 on a rampage formula and a q6600 at 3.6ghz system ran fine for months but temps got worse and worse so i sold my old setup and built a corei7 like yourself, but never addressed the problem, But for example now with my current build i went with a corei7 920, Asrock x58 supercomputer MB, 3x2gb Corsair Dominator at 1600, and a ABS Tagen 1100W PSU, system runs fine with one 4870x2 at 3.8ghz stable i can even post at 4.2 but crashes soon after in windows (windows 7 x64 build 7100) but as soon as i quadfire i have stability issues, vantage crashes during cpu runs!!! which makes no sense because on the both gpu test i easily get over 100fps with no issues so i did some air flow upgrades mind you that i run a complete custom water setup that consist of Swiftech gtz cpu block, Swiftech NB block, one DangerDen 2x120mm rad up front and a second 120mm rad at back with a push pull setup, a Swiftech highflow res rev2 with a Switftech MCP655 pump pushing everything through 1/2 tubing, all inside a Antec 900, a pretty good airflowing case but with all that still have crashes with quadfire, so my airflow upgrades consisted of spot fans to help with airflow but still nothing, it seems 4870x2's cause more heat than you think because im coming to the conclusion that the motherboard is just conducting the heat upward through the pcie slots into the SB and possibly the NB and causing the crashes, because heat causes resistance and resistance leads to processing delay's then bam you got a BSOD, i see that your board has pcie slots close like my rampage, but now my cards sit farther apart thanks to the board i have but yet agian heat is a bitch so im selling my quadfire setup to someone who wants to burn and going with a single gtx295 and forget ati all together, so brother ati isnt worth it anymore take my word for it!!! and if you think im crazy then try it you'll never look back