GA-990FXA-UD3 random reboot with FX8320

[simotion]

Hi there,

I've upgraded a system with a GA-990FXA-UD3 motherboard and a FX8320 processor, specs;

- Case: NZXT Hades
- PSU: OCZ Fatality 550w
- MB: GA-990FXA-UD3 rev 4.0
- CPU: AMD FX8320 (8 core, 3.5 GHz)
- Cooler: Scythe Grand Kama Cross
- Mem: 8 GB DDR3 (2x2 GB Vengeance, 2x2 Kingston Value)
- GPU: Asus GTX770 DCII

Other than that it has a Samsung SSD, a 7200rpm HDD for storage of games and a bluray optical drive, but I don't think that's relevant to the problem at hand.

The computer runs smooth  but when stressed or during gaming it just resets itself. On reboot the event log shows Kernel Power 41 (63) error.

I've plugged the pc directly into the wall socket and connected it to a current measuring device, it reports max power draw is ~473 watts when fully stressed by running Prime 95, heaven 3D and Furmark at the same time.

HWMonitor reports that the voltages are good (12v = 12,10 and 5v = 5.04) when stressed.

When I put the case's built in fancontroller to max, it runs stable for hours, but when I turn the fans even a bit lower it crashes within minutes (if I start stressing it with the fans on low it takes about 15 minutes to half an hour before the system resets itself).

HWMonitor gives the following temperature readings after an hour of stressing with all the casefans on highest speed:

- SYS: 51.4 C
- CPU: 60.2 C
- TMPIN2: 56.1 C

- Package temperature 51.9 C

The system isn't overclocked yet, everything runs at default speeds...

I've now swapped the FX8320 for a X6 1035T with the stock cooler, the system is running for two hours stable now with all the fans on lowest speed. This gives me the following temperatures;

- SYS: 36 C
- CPU: 58 C
- TMPIN2: 40 C

- CPU Core temperature is between 39.5 and 40.5 degrees across all 6 cores.

Maximum power consumption at the wall is about 50 watts lower with the X6 in the socket.

It seems that the FX8320 runs a bit hotter than the 1035T, wich I expected... But especialy the NB/VRM temperature and system temperature are extremely high with the FX8320 in place. With the FX8320 the heatsink of the NB and VRM's gets so hot that I burned my fingers on it. Is this causing the system to randomly reset itself without a bluescreen or any single hint of what's going wrong? I thought this board should be able to handle the FX8320, especialy at stock speeds, but it seems the boards VRM / NB cooling setup is having a real hard time keeping up with the 8 core... but although the temps are 'not cool', it should still be within limits of what it can handle I would think?

The NZXT hades has an exhaust fan both at the back and the top of the case, right where the VRM's heatsink is and the Grand Kama Cross is a top down cooler on which the fin's of the heatsink are angled in such a way that the VRM's and memory get a breeze as well (see picture).



I've also tried the stock cooler  but the only difference that this made was that the CPU ran 10 C hotter than with the GKC.

Has anybody any idea what's causing this and better yet, how to solve it? It's driving me quite insane.

[dmdilks]

Two things how much paste are you using. You need only enough to really fill in the cracks in the two metals.

The temps at idle should be between 30c to 40c. The 2nd you could have a bad CPU.

[simotion]

I put a pea size drop of thermal paste in the middle of the cpu before I placed the cooler on top of it. The same as I always do When I removed the cooler the spread looked good and thin.

Idle temps;

SYS: 31 C
CPU: 35 C
TMPIN2: 25 C

Package: 26 C

I can try the CPU on a Asus m5a78L-m/usb3, but since that board is 4+2 phase power if I'm not mistaking, and the cpu is already overheating the UD3, I'm afraid to put it on the poor ol' Asus... It could pretty well be the end of the Asus board, if the CPU is the culprit....

[Tiger]

Mixing memory kits together may cause compatibility issues such as unable to boot or unable to operate at rated specifications. Try using a single kit and enable the XMP (eXtreme Memory Profile) in the BIOS. Using multiple memory kits may need the memory clock adjusted to a slower speed to work.

[simotion]

The memory modules seem to get along just fine with the X6 1035T...

Other than that, the mem speed is set at 1333 MHz and timings are loosened up further from stock.

Memtest86 shows no error after running for 6 hours.

To be absolutely sure, I will take them out and install a dualchannel 8 GB vengeance kit I have laying around.... however to be honest I don't expect this to be the cause of the mysterious resets and high temperatures.

[dmdilks]

I put a pea size drop of thermal paste in the middle of the cpu before I placed the cooler on top of it. The same as I always do When I removed the cooler the spread looked good and thin.

I was only ask because I seen people put to much on and be all over the place. I use a BB size in the middle.

I still think you have a bad CPU. Very rare when that happens, but things like that do happen.

[AgentFXA]

Perhaps simply up the cpu temperature threshold in the system bios? - the default is 60C as far as I recall hence the system goes down. Set to 70C and all should be well!?

The X6 is a lame dog compared to the FX83xx and just make sure the side fan blows (unobstructed) air onto the mobo - this is especially important in hot (summer) countries.



 

[CalTek]

Just wondering if you found a solution?

I just built my system last night with the same processor and am having the same issues.

I would be in windows setting up an app and it would reboot. The weird thing is I would then be able to get in and play Battlefield 4 or Call of Duty on Ultra settings and it was fine then it would reboot when I was setting up a new Windows user.

Initial thought was maybe an issue with the memory as the motherboard wasn't initially picking up the 1866.

I am using stock fan on the CPU and will install monitor app tonight to see what temp gets too -- my thought is if it plays COD and BF4 without re-booting it shouldn't be a temperature issue.

[rinkol]

One question would concern the CPU fan speed - unless set otherwise in the bios, this should increase in speed as the CPU gets hotter.   You could get a temperature monitoring program such as Core Temp and get a better idea if there is definitely a temperature issue. Furthermore, if the temperature gets dangerously high, you should expect the CPU fan to speed up substantially.

I ran into a strange bug with the F3i bios in the same board. On waking from sleep in windows 8.1, the CPU fan would not run or run slowly and a crash would soon follow.

Robert

Hi there,

I've upgraded a system with a GA-990FXA-UD3 motherboard and a FX8320 processor, specs;

- Case: NZXT Hades
- PSU: OCZ Fatality 550w
- MB: GA-990FXA-UD3 rev 4.0
- CPU: AMD FX8320 (8 core, 3.5 GHz)
- Cooler: Scythe Grand Kama Cross
- Mem: 8 GB DDR3 (2x2 GB Vengeance, 2x2 Kingston Value)
- GPU: Asus GTX770 DCII

Other than that it has a Samsung SSD, a 7200rpm HDD for storage of games and a bluray optical drive, but I don't think that's relevant to the problem at hand.

The computer runs smooth  but when stressed or during gaming it just resets itself. On reboot the event log shows Kernel Power 41 (63) error.

I've plugged the pc directly into the wall socket and connected it to a current measuring device, it reports max power draw is ~473 watts when fully stressed by running Prime 95, heaven 3D and Furmark at the same time.

HWMonitor reports that the voltages are good (12v = 12,10 and 5v = 5.04) when stressed.

When I put the case's built in fancontroller to max, it runs stable for hours, but when I turn the fans even a bit lower it crashes within minutes (if I start stressing it with the fans on low it takes about 15 minutes to half an hour before the system resets itself).

HWMonitor gives the following temperature readings after an hour of stressing with all the casefans on highest speed:

- SYS: 51.4 C
- CPU: 60.2 C
- TMPIN2: 56.1 C

- Package temperature 51.9 C

The system isn't overclocked yet, everything runs at default speeds...

I've now swapped the FX8320 for a X6 1035T with the stock cooler, the system is running for two hours stable now with all the fans on lowest speed. This gives me the following temperatures;

- SYS: 36 C
- CPU: 58 C
- TMPIN2: 40 C

- CPU Core temperature is between 39.5 and 40.5 degrees across all 6 cores.

Maximum power consumption at the wall is about 50 watts lower with the X6 in the socket.

It seems that the FX8320 runs a bit hotter than the 1035T, wich I expected... But especialy the NB/VRM temperature and system temperature are extremely high with the FX8320 in place. With the FX8320 the heatsink of the NB and VRM's gets so hot that I burned my fingers on it. Is this causing the system to randomly reset itself without a bluescreen or any single hint of what's going wrong? I thought this board should be able to handle the FX8320, especialy at stock speeds, but it seems the boards VRM / NB cooling setup is having a real hard time keeping up with the 8 core... but although the temps are 'not cool', it should still be within limits of what it can handle I would think?

The NZXT hades has an exhaust fan both at the back and the top of the case, right where the VRM's heatsink is and the Grand Kama Cross is a top down cooler on which the fin's of the heatsink are angled in such a way that the VRM's and memory get a breeze as well (see picture).



I've also tried the stock cooler  but the only difference that this made was that the CPU ran 10 C hotter than with the GKC.

Has anybody any idea what's causing this and better yet, how to solve it? It's driving me quite insane.

[AgentFXA]

Low fan speed after sleep:

Apparently Gigabyte hastily released a hardware revision update for my 4.0 board (now 4.1) which seems to suggest it's a design flaw.

[CalTek]

Low fan speed after sleep:

Apparently Gigabyte hastily released a hardware revision update for my 4.0 board (now 4.1) which seems to suggest it's a design flaw.

Interesting as my reboots happen when I am in the middle of using Windows or just logging in. I will look into this update.

To date I have done the following to figure out the reboot issue:

• swapped out my 16 GB DDR3 1866 sticks with older 2x2GB DDR 1066
• unplugged all extra HDD, fans, USB to see if PSU (600watt) was issue
• reinstalled different version of Windows

All tests still resulted in an unexpected reboot. I ran a Memtest and the RAM was ok as well.

I found that on my end things were fine when I was installing the Windows 7 and were fine until I updated the video drivers and went dual screen. Once in dual screen that is when the issue seems to happen. This lead me to believe my issue may have been my video card (Zotac 960GTX). I decided to switched over to single monitor to see if it was the dual monitor situation that was the issue (main montior - mon. #1 is a 24" LG and mon #2 is an older 19" BENQ) and was able to use it for an hour without issue.

Initially I was running mon #1 in HDMI mode and the mon #2 in DVI but now switched mon #1 to DVI and and mon #2 to a DVI-to-HDMI conntection (Zotac only has one DVI port and one HDMI). After doing that it ran last night and when I woke up this AM it was still on (mind you it was in sleep mode - forgot to disable).

This morning I disabled sleep mode and have it running while I am at work. I will check when I get home and will update on status tonight or over the weekend as I plan on gaming and installing more apps tonight.

I did update the BIOS via @BIOS app in Windows so maybe that did something and I think I may have disabled Turbo mode on the board so maybe that was the fix.

Regardless I have 14 days to figure out if I am going to keep or return the board.  After this first impression I may return it Sunday for a different brand - especially if it reboots any more over the weekend.

I love Gigabyte but this is frustrating and annoying and if I have this many issues now how many will I have in a year or two?