Help with my computer #165468161
Hey, long time reader in need of help.
Been having problems with my computer freezing while playing SOME games. I can manage games such as Deus Ex: HR, Portal 2 etc fine but more intense games such as CODMW2 and BF3 end up freezing after 10-15min. Its not getting hot (45C loaded) and I'm running the latest Catalyst drivers.
The system is as follows:
2500k processor + Hyper 212
G. Skill DDR3 1600Mhz
Asus P8Z68-V LE
Sapphire HD6870
500W PSU
Any construtive help would be appreciated.
|
copuis
Brisbane, Queensland
1709 posts
cpu might not be getting hot, but what about your video card???
and what brand of psu, because a cheaper one might have issues (but a 500w should manage, not ideal, but manage)
and what sort of freeze?, is it a picture displayed, and the whole she-bang stops for a bit (that might point to a hard drive that is not spinning up in time, or simply out of space when caching the files)
is it like a power drop, screen goes blank, then it goes back into the game, or crashes, that could be a video card in thermal distress
|
Sc00bs
Brisbane, Queensland
8307 posts
if its just freezing/crashing the game in bf3, then your in the same boat as alot of others =(
|
The video card sits around 50C idle, up to around 70-75C while gaming. I've monitored the fan speed and on the auto setting it doesn't really get over 55% capacity, so I manually set it at 65% (fucking loud) and the temps hovered between 65C-70C but still crashed.
It seems like a full system crash. Can't tab out at all. The video just freezes and the sound goes epic loud and shudders.
|
m3nt4l
Brisbane, Queensland
1730 posts
Have you opened it up and dusted recently?
|
copuis
Brisbane, Queensland
1716 posts
70-75 surely has to be closing in on the shut down temp set the fan to really fucking loud (80%) and see if that helps
|
Its only recently new... less than 3 months old now.
I've run the quick Windows 7 memory test and it was clean... probably should do a proper memtest over night right?
Funnily enough, I had a crack at Prime95 just now and it crashed after a few minutes... just froze up and didn't respond which makes me think its not the video card.... fuck
|
copius, tried it running at 75% and it didn't get over 62C and froze... rage!
|
m3nt4l
Brisbane, Queensland
1732 posts
I clean mine out about every three months, you'd be surprised how much shit gets stuck in the heatsinks :(
|
copuis
Brisbane, Queensland
1718 posts
Tiny
Brisbane, Queensland
3012 posts
75 degrees seems a bit hot to me.
|
I'll try that tonight.
I'll mention that when benchmarking in Heaven DX11 it runs well and doesn't freeze and thats basically GPU loading only.
|
the only time i've had issues with some games and not others it turned out to be a RAM issue (voltages weren't set correctly in BIOS).
maybe run a memtest?
|
slyin
Brisbane, Queensland
110 posts
memtest memtest memtest if you havent
Enjoy your testing..
i just went through a similar thing with my similar system. for me it ended up being dropping my ram voltage down to 1.5v instead of the recommended 1.625v, kingston though. it had no problems making it through a night of memtest with no errors but would lock in prime. the amount of testing i did was making me v.angry, so once again i say enjoy urself.
im not saying this is your problem as i have no idea what state your system is in. just venting really
last edited by slyin at 17:42:50 02/Dec/11
|
simul
Brisbane, Queensland
1256 posts
Make sure the cpu fan is properly seated - a month ago I was having basically the same issue - played with the fan and thought it was on right, so I assumed power supply and swapped it over. But turns out while it felt secure one of the clips had popped out and that was enough to make it overheat to shit.
Not saying thats the issue, but it definitely was to me and wasn't until I really really checked that I noticed (past cpu/mobos I've used you could easily tell that the fan was loose, but not on later ones).
|
Nukleuz
Perth, Western Australia
107 posts
Without knowing your full system specs it's hard to diagnose but it is coming in to summer. Weather is getting warmer and PSU's are definitely less efficient once it gets hotter and is under load.
Given it takes 10-15 minutes for a crash I do suspect it could be your PSU (voltage drops causing instability), you have a fan that is not seated properly, or lastly you got unlucky and you have faulty RAM. Worse case scenario is it is a faulty mainboard although being an ASUS that is a lot less common.
Things you can do:-
1. Check memory timings, voltages etc. If in doubt let it go auto (by SPD)
2. Reseat the RAM.
3. Take half the RAM out - is it then stable? If not replace the stick you took out and take out the other half. Retest.
4. Make sure all the latest drivers are installed.
5. Update the mainboard BIOS.
6. Try a different power supply (higher voltage - 650 minimum and a reputable brand name ie Corsair, Thortech, Antec, even Thermaltake are pretty good)
Additionally CPU-Z and GPU-Z are useful. |
Hey guys, thanks for your replies.
I've just finished 4 hours of memtest and nothing. Since then I've pulled the covers off and rechecked everything to make sure nothing n00b had happen and it all appears fine.
The rig was put together by the Umart mob in Southport cause I CBF at the time and its better off warranty wise anyway. Case has 4 x 120mm fans moving shitloads of air and according to the mainboard sensor, temps don't get over 35C and CPU hasnt gone over 45C. Its doing my head in!
Full Specs shamelessly copy\pasted:
Noctua 120mm 1300RPM P-Fan
Cooler Master Hyper 212 Plus
OCZ 120G Vertex II E Series SSD
CoolerMaster RC430 Black Edition 500WPSU
Sapphire HD6870
Seagate SATA3 2TB Barracuda
G Skill 8G(2x4G)DDR3 1600Mhz
Intel Core i5 2500K
Benq GL2440HM 24"W
Asus P8Z68-V LE 1155 Socket
Running Windows Ultimate x64 Service Pack 1 and everything is at stock specs.
Also slyin, RAM is at 1.5V atm :/
|
Reverend Evil™
Wynnum, Queensland
19469 posts
Prolly the shitty 500w PSU
|
parabol
Brisbane, Queensland
6634 posts
OCZ 120G Vertex II E Series SSD
Ah, yes the crashorama Vertex 2 series ...
|
Nukleuz
Perth, Western Australia
108 posts
Ah, yes the crashorama Vertex 2 series ...
I assume you are referencing the sandforce bug of which there is a BIOS update which stamps it out.
That's another thing for you to check Bromby.
|
Spook
Brisbane, Queensland
33634 posts
OCZ 120G Vertex II E Series SSD
Ah, yes the crashorama Vertex 2 series ...
so you keep saying.
i havent had a single problem with mine
i put them in a fair few machines ive built, and none of them have had any problems either.
the only ssd issues ive had were with series 3's ocz's and corsair force series 3
|
Sweet I'll check the hdd state out. I ran Prime95 out all last night, didn't get over 55C. Rah!
|
step
Brisbane, Queensland
2369 posts
The Vertex 2's were fine, it's the 3's that have issues.
|
copuis
Brisbane, Queensland
1726 posts
hows it running now? (i guessing you've done some cleaning and re-seating)
|
Hey guys, well I've updated the mobo to the latest BIOS, updated the firmware for the SSD and clean and reseated everything bar the CPU.
It was fine for an hour or so and I though I had it beat, but just now its frozen again :/
I was keeping an eye on the temps and it froze at 45C CPU and 67C video temp. Bullshit!
Just a curveball question but any chance that the wireless card (not previously mentioned) would conflict with anything? Its a D-Link DWA 525 and from what I've read they're a tad bit shit, though not much surrounding crashes. Grasping at straws here. I would try another PSU but I don't have one. I'm gonna do a roll back on the Catalyst drivers later and see what happens.
|
copuis
Brisbane, Queensland
1728 posts
man I would think that once the temps on the Video card start getting and holding above 55c you're asking for an issue
pull the side off, stick a big fan blowing air in, check if that helps, if it does, ditch the video card/cooler, and replace
|
funny thing is copuis that it only SOMETIMES freezes under stress load with Prime95. Prime95 shouldn't affect the video card at all and the reviews\tests I've researched subject that sub 70C temps are normal and pos 90C should be discouraged... shit balls fuck
|
Pinky
Melbourne, Victoria
11636 posts
Edit: In order to provide something constructive have you tried ripping out all except one of the RAM boards then run Prime95? If that fails, take that one out, put in another and try again? See if you can isolate one of the RAM modules as the fault.
|
Tollaz0r!
Brisbane, Queensland
11991 posts
Check your PSU, on the side of it what is the rated amps for the 12v rail(s).
If you have more then 1 rail make sure you have put your CPU and GPU on seperate rails. With your setup I would think you may need 300w your combined 12volt rails, with minium of 150 to each rail.
Cheap PSU's often say they are 500Watt, maybe all the rials combined, however they usually peak out at 350 or something like that on the 12v rail with little else for the others.
|
skythra
Brisbane, Queensland
4920 posts
Prolly the shitty 500w PSU
funny thing is copuis that it only SOMETIMES freezes under stress load with Prime95. Prime95 shouldn't affect the video card at all and the reviews\tests I've researched subject that sub 70C temps are normal and pos 90C should be discouraged... shit balls fuck
This is a symptom of a bad psu, not that the psu is faulty, but the psu isn't able to keep up with the heavily dynamic changing load.
For example if your load needs to be at 80% of power capacity, during a benchmark it might get there at a predictable range and then it stays there. Amps and volts meet the requirements, but if it's in a game, which isn't predictable for needing 100% usage or 80% or 20% then the amp changes fluctuate to reach the goal, but the voltage may be diminished or even brown out causing those issues.
I'd say borrow a friends PSU if they have something closer to 875watt and swap it in (without sli this should be simple if your case is btx, or if it's atx you might need to remove your cpu cooler).
because of dynamic changes you should always overspec your PSU by at least 20% of what you think the max will be, because to reach it quickly in time to process requests, it needs to easily surpass the amp/volt draw of a computed calculation.
|
Tollaz0r I'll check it out when I get home and report back!
Skythra and Pinky, thanks for the info mate. I'll redo the memstest stick by stick and Prime95 the balls out of both of them and if it's still fucking up I'll buy a bigger PSU. Something like this |
Whoop
Brisbane, Queensland
18993 posts
man I would think that once the temps on the Video card start getting and holding above 55c you're asking for an issue
pull the side off, stick a big fan blowing air in, check if that helps, if it does, ditch the video card/cooler, and replace
my 285 used to run over 80 degrees just fine, the CPU in my MBP runs at 100 degrees. Why are people so damn paranoid about temperatures these days?
|
step
Brisbane, Queensland
2370 posts
man I would think that once the temps on the Video card start getting and holding above 55c you're asking for an issue
lol
funny thing is copuis that it only SOMETIMES freezes under stress load with Prime95.
Are you overclocking? If youre PC is failing prime95 at any point, you dont' have a stable overclock. All this extra testing is useless if this is the case.
|
DK
Brisbane, Queensland
531 posts
Temps are fine. Sounds like psu to me.
|
copuis
Brisbane, Queensland
1729 posts
whoop, massive changes in temps are long term bad, chip creep is not as likely, failure is expand and contract enough and it starts to stress parts out
and when my gts450 get to 55 it starts to fail, now I know one the other factor might be that the temp probe is not on the hottest part so the true temp might be higher, but that is going to be the same on any bit of hardware
also, sticking a fan in the side is a cheap, easy fault finding method, and once done you can pretty much cross heat off the list of possible causes
|
step, everything is at stock settings
If you have more then 1 rail make sure you have put your CPU and GPU on seperate rails. With your setup I would think you may need 300w your combined 12volt rails, with minium of 150 to each rail.
PSU is rated to 360W combined 12V rails. Just getting home now so I'll check if the CPU and GPU are on separate rails. |
Looks as if its a fault stick of RAM. After hours of memtesting each stick with no errors, I fired it up one stick at a time. Just had my first decent go at it for about an hour with zero problems. Fuck you RAM!
Thanks guys for the feedback, off to Umart tomorrow for an exchange yall
|
jmr
Brisbane, Queensland
7501 posts
An exchange
hahahahahaha
I don't think you understand
|
I've taken ram back for them before and wasn't a problem. Straight in straight out at Milton store. Unluckiest person with le rams
|
|
PM
WWW