Persistent GSOD and BSOD :(

k02033

Distinguished
Aug 12, 2010
16
0
18,510
I brought a brand new $1300 computer that constantly experiences GSOD and/or BSOD whatever games i play. The list of games i have tried includes: Modern warfare 1 and 2,

bioshock 1 & 2, bad company 2, starcraft 2, warcraft 3, deadspace, need for speed shift.

============

More explanation of the crash:

- first the game screen would pause, then the sounds of the game would loop for about 5 secs then a solid grey screen appears. After this the monitor will loose signal from the computer and the system is no longer responding and requires a hard restart.

- sometime there is no grey screen and a BSOD occurs

- most of the time the crash occurs within 10 mins of game play. Sometimes oddly enough it can run for ~2 hrs without crashing.

============

I am using:

Sapphire PCIE ATI HD5770 Vapor-X 1GB 128-bit DDR5, 2xDVI, HDMI, DisplayPort

AMD Phenom II X4 955 Black Edition Quad Core 3.2GHz 8MB Cache AM3/AM2+/AM2 CPU

ASUS M4A87TD EVO AMD 4xDDR3, 2xPCIE, RAID

1TB 1000GB Samsung Spinpoint F3 HD103SJ SATA II 3.0Gb/s 7200RPM 32MB Cache

4GB DDR3 G.Skill 1333MHz F3-10600CL9D-4GBNT CL9-9-9-24 (2x2GB)

23" Samsung Lavender P2350

450W Corsair VX-450 Power Supply

windows 7 64 bit

============

What i have tried in attempt to fix this:

- Windows and games are all fully updated

- Furmark and GPUz shows the average temperature while gaming is around 65 degrees, so its not a heat problem

- I tried ATI's 10.1 10.1 (hotfix) 10.2 10.3 10.4 10.5 10.6 10.7 10.7a beta and sapphire's 10.7 drivers. None of these solve the problem

- I turned windows themes off, and all the aero effects are off too

- I made use of driver sweeper in safe mode to make clean installs

- I tried all permutations of the game settings including various resolutions, this didnt fix anything either

- Bios is up to date and all bios settings are factory defaults

- I tested the computer with Prime95 for 8 hrs and also ran memtest all without errors, meaning CPU and RAM are in good health (is this a valid conclusion?)

- I have formatted windows numerous times

- I have not installed any unnessary programs, i completely disable anti virus when i test the computer with a game.

- I have "high performance" set for my power plan

- I have not overclocked any components

- This is not a dust issue since its a new pc

- The computer shop i brought the pc from tested a brand new video card in place of mine and the same crashing occurs (although i did not see them do the test personally, so not 100% sure on this)

- However I am able to run furmark's stability test for 2hrs without issues

- the last two points seem to suggest that RMA will not solve my issue

============

I am quickly running out of ideas :(

Any help/ suggestions are much appreciated
 

Griffolion

Distinguished
May 28, 2009
1,806
0
19,960
Have you manually set the voltages and timings for your memory in the BIOS?

Often motherboards setting them to 'auto' isn't actually correct for the RAM and manually setting such things is often needed. Undervolting memory (which is often what the auto setting does) can result in problems like what you have said.

I would advise you do this and then run Memtest on your sticks to see if there is a faulty memory issue.

If possible it might be an idea to run chkdsk as well to make sure it isn't your HD.

Try those for now.
 

k02033

Distinguished
Aug 12, 2010
16
0
18,510


I think you may just hit the solution, 90% of mine bios setting are on [auto] and this is the factory default settings. Could you link me to a novice friendly guide on bios settings? That would be a great help. Thank you.

Oh and i have ran memtest, results show no errors but i will still look into the bios voltage settings, seems to be plausible fix.

And i will try chkdsk as well.

 

k02033

Distinguished
Aug 12, 2010
16
0
18,510


Heres the blue screen code:



Problem signature:
Problem Event Name: BlueScreen
OS Version: 6.1.7600.2.0.0.768.3
Locale ID: 3081

Additional information about the problem:
BCCode: 116
BCP1: FFFFFA8004085310
BCP2: FFFFF88003E0CC80
BCP3: 0000000000000000
BCP4: 0000000000000002
OS Version: 6_1_7600
Product: 768_1

Files that help describe the problem:
C:\Windows\Minidump\080310-15412-01.dmp
C:\Users\\AppData\Local\Temp\WER-30186-0.sysdata.xml

Read our privacy statement online:
http://go.microsoft.com/fwlink/?linkid=104288&clcid=0x0409

If the online privacy statement is not available, please read our privacy statement offline:
C:\Windows\system32\en-US\erofflps.txt
 

k02033

Distinguished
Aug 12, 2010
16
0
18,510
I'm pretty sure 65c is the limit for your processor, either cool the room better or get some better case/cpu cooling.

Not saying this is the cause of your problem, but it needs to be looked into.


65c was the temp for my vid card. I am not sure but vid card can handle much higher temps than this right?



Oh and i shut off defender, i was able to play a modern warfare 2 match without crashing (10mins) then i crashed on a map change. However the crash was different. This time it was just a blank screen, no sound loops no grey screen no BSOD but the system still locks up and i had to hard restart. Um i will try turning off window firewall as well now. will report back in ~30mins.
 

k02033

Distinguished
Aug 12, 2010
16
0
18,510
Ok this is odd. When i disabled windows update service and windows defender service i was able to play modern warfare 2 and starcraft 2 on the highest settings for ~2hrs straight without any troubles at all. But when i exit out of the game and started playing again the crash would occur within 5mins again. Its as if some parts of windows is interfering with my video card. Is everyone familiar with this behavior?
 

Griffolion

Distinguished
May 28, 2009
1,806
0
19,960
How did chkdsk go?

Also, in the BIOS you will want to be setting the DDR or DDR3 voltage to your RAM sticks recommended amount (standard is 1.5v but some can go to 1.65v, check your own sticks). The timing settings will be in memory calibration, if you just explore your BIOS you should come to the memory timings section. The memory timings given for your sticks should go as follows (in numbers): X - X - X - XX (where XX would equal a 2-figured number). My timings for my sticks are 8-8-8-24, yours will be something similar to that.

In the BIOS the first number represents "CAS" (aka tCL), the second represents tRCD, the third represents tRP and the last one represents tRAS. Simply enter the timings in that order, so using the example i gave above for my timings it would be as follows:

CAS - 8
tRCD - 8
tRP - 8
tRAS - 24

Hope this helps.
 

Griffolion

Distinguished
May 28, 2009
1,806
0
19,960
Again, sorry for the double post, but toms doesn't let me edit my posts!

Anyway, after reviewing your first post in the parts section, the timings for your memory sticks are 9-9-9-24, that should save you having to check the side of your sticks.
 

Griffolion

Distinguished
May 28, 2009
1,806
0
19,960
No problem. Just to clarify, in the timings section of the BIOS you may see many more types of timings listed that i havent listed (could be up to about 30 different types). Leave these on auto, the four i have gone through here are the main ones you need to set.
 

k02033

Distinguished
Aug 12, 2010
16
0
18,510
Um when i want to change the timing values the options are the numerical value along with "clk" i.e. they look like [9 clk], what does the clk mean?

I also found that my ram should work under 1.5V
 

Griffolion

Distinguished
May 28, 2009
1,806
0
19,960
clk will most likely mean 'clock'. It is used in this way to represent the amount of milliseconds in every cycle for that particular function. Just set them how they should be.

You will need to find out at exactly what voltage your memory should be running at and set it to that.
 

k02033

Distinguished
Aug 12, 2010
16
0
18,510



Yep the memory should be running at 1.5V. I might as well take the time to set the other values in the bios as well. We have to consider the voltages of the CPU and the video card itself right?
 

Griffolion

Distinguished
May 28, 2009
1,806
0
19,960
They are usually alright when set to auto, its just the memory that can be a bit funny sometimes due to SPD standards for DDR3 being 1.5v at stock but some stick manufacturers seek out extra performance through putting higher voltages through the sticks, hence the extra need to make sure you get the memory settings right.

Everything else on auto for the time being will be fine.
 

k02033

Distinguished
Aug 12, 2010
16
0
18,510
Hi,

I have set the voltage and timing for the ram, but sadly it didnt work, the crash still occues. But i learnt a little about the bios, which is cool. Thanks!

And heres an update:


Last night i disabled lots of unnecessary windows services, including windows theme, aero effects, windows update, window defender, media player, action center, and other unnecessary services. Immediately after doing this, i launched modern warfare 2, and i was able to play ~1hr worth of very graphically intensive missions without any issues what so ever. The frame rates were good, loading times were good, no stutters, no troubles. I then loaded up sc2 and was again able to play it on the highest settings for ~2hrs without problems. Heres the weird part: I quit sc2, and then relaunch the game and it crashes pretty much immediately, i re launch modern warfare 2, and it too crashed very soon. I have not made any changes between those 2 periods. This is just bizarre to say the least. But i have a feeling that this is not some kind of freak accicident. Because in my early attempts at fixing the crashing, i disabled AVAST anti virus and this allowed me to run game without problems, but when i quit the game and relaunch them the crash come back, even though i have not enable avast back on. At that time i thought of this as just a freak accident.

Further more, lots of people are suspicious of my PSU: 450W Corsair VX-450 Power Supply

==========

Reasons why the PSU might be at fault:

- Its provide 450W, which is the bare minimum for my 5770.

==========

Reasons why it might not be the PSU's fault:

- The technicians at my computer shop (maco technology) ensures me that the 450W PSU is able to run my system, heres their exact words:



"Here is the link that calculate the power. Your system should easily be able to run quietly and efficiently on the Corsair VX450W PSU
http://www.corsair.com/psufinder/re...5798&cpu_oc=False&cpu_oc_ex=False&vc_oc=False

The Corsair 450W is sufficient to run your system."


- also the reviews on this PSU suggest it can actually deliver 570W despite its specs, see here:

http://www.hardwaresecrets.com/article/Corsair-VX450W-Power-Supply-Review/540/10

- and also from the experience i decribed above, the PSU is able to power my system to play 2hrs straight without crashes, so this huge discrepency between crash times i.e. ~5mins vs 2hrs seems to suggests its not the PSU's fault. i.e the PSU should give out at consistent times.


Personally i am not sure which is true at the moment. Perhaps i should try running furmark and prime95 at the same time to stress the system, see if the PSU gives out. Or maybe i should downclock my vid card...

==========


I should also be suspicious of windows 7


Reasons why windows might be at fault:

- The crash was gone when i disabled lots of its services.

- This is what maco technology believes in, here are their exact words:

"As AMD forum said this is a common problem with Windows 7
http://forums.amd.com/game/messageview.cfm?catid=260&threadid=124747&highlight_key=y&keyword1=SAPPHIRE%20VAPOR-X%20HD5770"


Reasons why it might not be windows fault:

- If it really is window 7 's fault, then shouldnt my problem be wide spread? i.e. every single 5770 user out there would have experienced this, and this doesnt seem likely.


==========

What do you guys think?

And thanks everyone for your posts!
 

k02033

Distinguished
Aug 12, 2010
16
0
18,510
I suggest u reinstall windows.and if u have dual ram try running one ram at a time.

Yea i have reinstalled windows many times already, I will try that ram suggestion. Although i dont think its a ram issue, i ran memtest for hrs with no issues.