• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Computer reboots/freezes when gaming

D

demonx

Limp Gawd
Joined
Jan 16, 2007
Messages
130
Hello, so I have an unusual problem and this is pretty long but I'm hoping someone can give me an idea of what's going on here.
So I just upgraded my graphics to 2 radeon 4870's because I wanted something better then my previous cards. I currently have an e8500, intel dx48bt2 motherboard with the latest bios, 4gigs ocz ddr3-1600 ram and these 2 4870's with catalyst drivers 8.9... I overclocked my cpu to 3.8ghz with my previous video cards to match my RAM and motherboards SUPPORTED speed - 1600mhz fsb.

It took me awhile to get everything 100% stable, because I wasn't very familiar with overclocking when I started and it seems like a good majority of people who have this motherboard have troubles overclocking it. I finally had it stable after many hours of testing, reading, etc.

My problem is now that I upgraded my video cards, my pc reboots/freezes whenever I start gaming. Games like Crysis Warhead will reboot or freeze almost every time whenever I get to the Crysis or Nvidia intro screen. UT3 ran fine one game I played, but another time it froze on a loading screen. Stalker Clear Sky froze twice in mid game...

I've tried changing voltages on my ram up and down from their stock voltage (1.9v), tried loosening ram timings from 7-6-6-24 (default) to 7-7-7-20, tried upping my vcore from the default 1.325 to 1.35 and 1.375, tried changing northbridge from default 1.3v all the way up to 1.4v, tried changing fsb voltage from default of 1.2v up to 1.3v with no avail...

I put everything back factory default and no overclock and my pc seems to run fine.

Now I'm guessing it could be a heating issue... But then again, pc runs fine with no overclock! The room where my pc is in is pretty hot anyways. My motherboards pci-express slots are so close together that my first card is nearly touching the top of the 2nd card. It's even running 10c HOTTER on idle and load. It idles at a warm 67-70c and load is 80-85c, even with my fan running at 45%. My 2nd video card, with much better airflow, idles at 58c and load is approx 67c and even runs at a lower fan speed...

OR it could be a power issue but I have an Enermax Galaxy 1000w power supply. One thing that bugs me is Everest now doesn't display my voltages like it did with previous video card... I reinstalled Everest and it's working normally.

As I'm typing this out, my cpu is overclocked... My pc seems stable overclocked until I start gaming.

Is it the video cards? I'm not sure, because with no OC it runs fine... Anyone got an idea what could be causing this? Thanks
 
what you're describing does sound power related. however, when it reboots are you getting a brief flash of a bsod? do you have halt on bsod turned on?
 
Well, usually what I've been doing is just trying to play Crysis Warhead because that game reboots/freezes much faster then other games. 90% of the time, the game will reboot once it hits the crysis intro screen. Sometimes I can play the game maxed out for 15-20 minutes before it just freezes. I get no blue screen or errors or anything - even if it just reboots.

Now what I don't understand is how come I can run everything fine at stock settings? That's why it's almost hard to believe it's a power issue.. But does anyone really know?

I can't stop thinking it's a voltage setting... But maybe it is a power issue. I'm almost tempted to RMA these and get one 4870x2 but I really want to use the xfire support... Pretty much why I bought this mobo in the first place.
 
Yea true... I thought I had a decent psu though... Enermax... 1000w.. Is that not enough for these cards or what?
 
demonx said:
Yea true... I thought I had a decent psu though... Enermax... 1000w.. Is that not enough for these cards or what?
Click to expand...

Yes, and that is a nice quality PS... I would look at your cpu voltage, is it geting enough volts?
 
run coretemp and check your temps on your processor and post them back.. so that we can see if the temps are your issue.. but im guessing its the overclock on your cpu if its freezing and rebooting.. usually you will get a bsod if its the gfx cards.. but if its running fine stock.. ill most likely bet its your overclock thats the issue.. 1000w psu is plenty fine for that setup.. unless your running 15 hdd's with a raid card as well..

are you running 4x1gb sticks of ram or 2x2gb?
 
I have a Xigamatek (spelling?) heatsink. My cpu idles ~ 40c overclocked AND on stock and full load doesn't exceed ~55c, both overclocked and stock... So I know my cpu isn't overheating. I run it when I game and I have a Logitech G15 keyboard so I can see the temps on the monitor my keyboard has.

I just turned my computer on and got a bluescreen of death... Gave me like a IRQ NOT LESS THEN OR EQUAL error or something like that. I'm currently at 1.35v on my cpu.

CPU-Z shows my core VID as 1.225v and that never changes... Don't know what's goin on there. Everest also says its just 1.3v but my BIOS is set on 1.35v... I wonder if I need to set some shit higher in the BIOS?

I have 2x2gb sticks and they are duel channeled.

Everest shows my 12v rail getting 12v, 5v getting 4.9v, 3.3v getting 3.2v and 1.5v getting 1.3v
 
next time the bsod comes on write down the error code.. set windows to not automaticly reboot on error.. there will be 3 series of numbers.. usually 0x00000124 0x000003425 0x00002134 something like that.. write those down post em or google them you will get the error its giving.. but ive usually found when getting that error its memory related.. which could be the issue that for some odd reason the memories not stable.. but that code will give you the exact reason windows failed..

if it is memory related id suggest maybe moving your memory to the other slots.. ive seen some boards that are quite picky about which slots they are in when overclocking.. this could possibly be your issue..
 
Yeah, I'll try that if it comes up. I'm still not sure what's causing this.. I was 20 hours Prime95 stable and had no problems like this before with my previous video cards.. :confused:
 
could be possible that it just wasnt loading the memory enough to cause the error if the cpu survived 20 hours of prime95..

since i havnt used an intel processor in almost 6 years.. cant remember does intel use an HT link between the pci-e slots and cpu? if so wonder if that could be the cause? or maybe taking 1 card out.. play crysis with the overclock and see if it happens.. if it does with a single card.. maybe swap the cards.. see if it happens again.. i dunno just coming up with theories at this point..
 
I appreciate what you're suggesting anyways, theories or not. I was planning on removing a card as well because this simply did not happen before the upgrade... Maybe it's something in the OS and I need to reformat? Idk, very doubtful but it could be a possibility.

I keep hoping it's something that can be fixed without RMAing or buying new crap. Such a pain doing an RMA... Maybe I can figure something out. If anyone else has any more suggestions, please post them. :cool:
 
if it works stock with everything i doubt you will have to rma.. its just trying to find where the problem is.. then trying to find a way to work around it..
 
I sure hope so but if I can't stabilize it I will probably RMA these cards and try one 4870x2. It doesn't make much sense how it's so unstable with these cards and my previous setup was 100% stable. Maybe because these cards require so much more power that it wants more/less from the other voltages? IDK

Before I had two 3870's. They were one slotters and only required 1 6-pin for power. I know I really didn't really need to upgrade but the cards are already way behind. One of the 4870's are way faster and better performing then both of those cards! A few games weren't up to my performance standard. I'm also a graphics/performance whore and I have some money to throw away. :p

I'm not home at the moment to try what's been suggested... Damn work.. I'll try raising some voltages although I don't know what to raise.. I've already tried so many combinations. Next thing I will try is removing one of the cards to see if it continues to happen.
 
btw i looked up your memory.. one thing i noticed the timing you used is a little off from the default timing.. it should be 1.9v at 7-6-6-24 the looser timing would be 1.9v at 7-7-7-24.

btw just to see if the problem really is power.. i did a check on the extreme power supply calculator lite thing using the info from your sig. and came up with:

Recommended PSU Wattage: 524w at 100% TDP on cpu 100% TDP on the entire system including your overclock. so i dont think your psu is the problem..

http://extreme.outervision.com/psucalculatorlite.jsp
 
There is no way that Enermax is the problem power wise. The way it could be the problem though is if its voltages are drooping. Get out your multi-meter and test some things.
 
Thanks sirmonkey. I filled out everything that I could on the site with 100% TDP and 100% peak load and it's saying recommended psu wattage is 608w.. There's about an extra 400w my psu should be able to support! I believe I can take the psu out of the picture...

It's got to be a voltage issue.. Or maybe the card is overheating.. But even at 95% fan speed, problem persists. 95% fan speed under load = ~75-85c... That shouldn't be normal but I also doubt it's causing the reboots/freeze because stock is fine! Argh

And my ram is currently set at 1.94v and 7-6-6-24 timings. When I first ran into my overclocking problem with my previous video cards, I could not get prime95 to be stable past 9 hours on one of the cores at 1.9v. I upped it to 1.94v and that's when I passed prime95 for 20 hours. Never had a problem since the video card swap. I have tried dropped it back down to 1.9v, even 1.86v but at 1600mhz, anything below 1.9v will cause problems.

to darkcow: I don't have a voltage meter. :\
 
95% fan speed on a 4870.. and 75-85C.. um yeah.. i dont think its speeding the fan up.. because at 95% it would sound like a WD raptor right next to your ear..

personally id say first thing set your memory to 1.95v@ 7-7-7-24 since thats all within warranty.. see if that fixes your issue.. if not then go on to removing 1 gfx card.. test it.. take that one out.. test the other one.. pain in the butt but i think this is the only way to narrow it down to figuring out the problem..


lol edit again.. i saw your review on your card you bought.. another thing you could do is take a piece of paper.. fold it up a few times and stick that between the cards.. should be able to get atleast a few more cm's of clearance.. *stole that idea from the quad sli 9800gx2 review that was done a while again.. :p*
 
The VID is a fixed property of the individual CPU and nothing to do with your actual voltage. The voltage you measure in Windows will always be lower than the one you set in the BIOS as the motherboard allows some room for fluctuations.

What are you using to read your CPU temps? Different utilities vary by quite a lot, so it's possible that if you're hitting high 50s in one utility, you're actually reaching 70+ degrees in Coretemp. It wouldn't hurt to check temps in Coretemps to confirm that's not the problem.

Also, I'm not an expert on DDR3 RAM but those timings look very tight fo 1600. Are you sure they're officially supported at that speed? Even if they are, I'd suggest loosening them by a lot to see if this fixes the problem. It's not like you'll notice any difference anyway.

There's not much point in RMAing at random when you don't even know what component is at fault at this point...
 
official timings for that memory is1.9v@ 7-6-6-24 @ 1600 with a looser timing of 1.9v@7-7-7-24 1600
 
I was gaming back at stock settings for about an hour with no problems... Automatic settings clock my ram at like ddr3-1110 at 555mhz and 7-7-7-20 timings and 1.54v which is odd.. I can't seem to get it stable at 666mhz (ddr3-1333) because I'm not sure of the timings. I've tried 1.9v, 1.86v, 1.7v.. nothing.. But the stock settings for the ram should be ddr3-1600 @ 800mhz, 7-6-6-24 timings and 1.9v like what's been said.

Another odd thing is my cpu still underclocks and changes the multiplier when it idles, even with C1E and speedstep disabled... Hmm...

dmfg: my cpu temps are fine. I use CoreTemp and it idles ~40c and load never goes past 58c. 20 hours of 100% load on both cores and the highest I saw it go was ~65c. I don't think it's a temperature issue.

Its possibly the ram or mobo.. or a voltage setting.. still testing
 
Yes sir, I do in fact have the latest bios, 1814. Had it the week it was out. :D

I just don't get it. I thought I had it working because I was able to play this damn new Crysis game for about 10 minutes then it froze. :( I was just messing with the fsb and northbridge voltages in order to get it running.. Now I'm thinking it's the ram. I'll try loosening them to 7-7-7-24 and dropping the ram to 1.9v. If that fails to work I'll probably be forced to remove one of the cards and see what happens.
 
Ok, well here's an update of what's going on..

I had to reformat my pc last night. Apparently removing one of the vcards without uninstalling the drivers isn't a good idea. A black box covered 99% of my monitor when windows booted up. Going into safe mode and removing the drivers with Drivercleaner didn't seem to completely remove everything... Couldn't reinstall the drivers... So i just decided to reformat.

Fully updated and installed Crysis again... Started playing with just one card. Everythings fine but about 45 minutes into playing the game my system reboot :(

Not really sure what's going on... I might load up memtest86 again so maybe I can pinpoint the problem...
 
I was having a similar problem, increasing the voltage to both my north and southbridge cleared it up...
 
holy crap i never had that happen to me before.. i take my 3650 out of my other computer all the time.. though its running in hybrid crossfire with my 780g motherboard..
 
Well, I believe I fixed it. I had to clock my memory from running 333/1333 to 400/1333 which clocks my memory at ddr3-1333 instead of 1600 like it used to be. Same timings and voltages as before.. I can't believe I didn't try that before. I have tried every other voltage combination that seemed logical but the one that actually works. :rolleyes:
Not sure why upgrading my video cards require it to be set this way to run normally but I don't really care since there's no huge difference... Just would like to know why.
 
good to see you atleast got it working.. makes me wonder if that memory really is ddr3 1600..
 
Thanks. According to the stickers on the RAM they are. It's very strange.. I really want to blame my motherboard for the troubles.

I'm still wondering why it's only stable this way with 4870's compared to 3870's..
 
You must log in or register to reply here.
Back
Top