Hey guys, my machine has a stability issue which I would like to isolate.
Intel Q9550 Core 2 quad -cpu
Gigabyte GA-EX58-DS4 - mobo
4870 x2 - Radeon (force3d) - GPU
8gb ddr2 -(afraid I dont have the ram info at hand)
ocz 900 watt psu
The machine will crash every now and then. Sometimes weeks can pass before a crash. Sometimes once a day. Far from frequent but aggravating when playing an important game of SC2 with friends.
The types of crashes I experience is the computer freezing and hanging for a period of time. The computer will rarely recover and display a message like "windows has recovered from an error" or "display driver has restarted"
Alot of the time it will go blue screen saying "commencing memory dump"
It mainly only happens during heavy load so my main suspect is the 4870 x2. When the computer was built, the graphics card died within the warranty so I was sent a new one. But this didn't resolve the crashing issue.
I also noticed that the card runs very hot (80 idle, 100 load) and I also got some graphic gliches.
So recently I revamped the cooling in my system by investing in a 'coolmaster haf x' case and the card has dropped in temperature but the crashes still happen. Not as frequent but the problem hasn't gone away. I have noticed less graphic glitches though.
Just to clarify that the computer isnt overclocked. I even tried underclocking the GPU.
I did however notice something. I recently tried a few guides in OC my CPU (only a lil past 3ghz) and the exact same crashes became more frequent.
So sorry for my rambling but was I barking up the wrong tree? Could my cpu actualy be the culprit to these crashes? Maybe Bios settings? Or def my GPU?
The machine still handles games fine and think it will do fine for another year or so.
I decided to go through the BIOS and disable anything which could throttle or lower cpu load if temps get too high.
15 minutes of FurMark and cpu stab test didn’t cause any issues.
However 5 minutes into crisis caused a crash. However this time I was fairly prepared. Had the cpu and gpu load and temps in a separate window during play.
Honestly, I don’t see how it could crash. There was nothing exiting going on in game. Me just about to open a jeep door.
Cpu load was on 65% and one of the cores was 60c rest at about 55
GPU was at about 80% load and 85c
Going to try updating the bios. Aside from that I am out of ideas
Just to mention which I forgot with the first post is that I have a fresh install of windows on a new hard drive. All drivers are of course up to date.
But I am using windows 7 home addition. I see a lot of people with ultimate but I wouldn’t see how home addition could be causing me issues.
Flashing the bios and putting a newer version was a scary moment. But all went to plan and my Mobo now has the latest version.
Did not fix any issues though. Still crashing when under a lot of load.
I decided to pull out the big guns. I was out of ideas and so I pulled out the 4870x2 and popped in a friends standard hd4870.
I had a go at butchering it with 10 runs of a crysis benchmark. Followed by a session on crysis with the same settings.
The strange thing is the frame rate 'felt' better. Even though the numbers said something else the standard hd4878 framerate felt better during gaming and of course, no crashes so far.
During the crysis benchmarks I decided to open up the 4870x2 to hopefully find the thing caked in dust inside. Disappointed to see that there was actually very little dust so I believe its just the unit. Just disapointed as its the first time I tried radeon and I have been through two 4870x2 cards. There has been some negativity online regarding the GPU also.
Of course its early days and I will try to stress test the machine as much as possible to try and cause a crash but I think a new rig is going to be ordered soon, starting with a GTX480
Set all your ram to mobo default speed if it has been adjusted otherwise.
Download Memtest86+http://www.memtest.org/Download - Pre-Compiled Bootable ISO (.zip)
Burn a disc
Load from cd in boot priority
Test each module/stick one at a time
if you have alot of errors on a stick that's the culprit
if all the ram passes you can rule it out as a cause
Right guys I beleve I found the issue.
I know it has been a while but I think its a good idea to post what I did incase any one else is having a similar issue.
Even with the new card the machine was still crashing however. No graphical gliches but the odd random crash, memory dump, followed by windows blaming the GPU.
So I started testing each memory stick but, still noticing no issues with memtest86. While I was checking each stick I did take a better look at the memory and the sticker said it was rated at 1.9 volts. I also noticed after poking around in the bios a couple hundred times that the voltage to the ram was slightly below that, 1.85 volts (or something like that).
So I bumped up the voltage a tad and after a month, not had a single crash.
It was a difficult issue as my machine was crashing very infrequently. Could be as frequent as twice in a day or once in a week. But it was still happening. Not a crash in a month is very rare.
But yeah think thats it, hopefully this may help anyone who runs into the same issue as me.