Sign in with
Sign up | Sign in
Your question

Game crash(within 30mins), system locks need hard reset

Last response: in Video Games
Share
September 25, 2011 4:07:11 PM

Hello, I've googled and read various forums on similar issues to mine and after trying many possible solutions, none of them changed the outcome at all. I do my best to detail out everything as much as I can.

System specs

This system is only 2-3weeks old and self assembled.

Processor: Intel i5-2500
Mobo: Asus P8P67-M-B3
RAM: 8gb G-Skill RipJawX 1600MHz
GPU: Asus ENGTX560 Directcu II 1GB
Sound: (on board) RealtekĀ® ALC 887-VD2 8-Channel High Definition Audio CODEC
PSU: Andyson F600M 600W Modular 80+
Casing: CM Elite 430 (with 1 more fan)
Monitor: LG IPS226
Internal HDD: Seagate 1TB 32MB
DVD drive: LG GH24NS50 24x DvD-Writer

OS: Windows 7 64-bit
DirectX: DirectX 11
Resolution: 1920x1080


The Problem

I first ran into the problem when I played StarCraft II. I run the game on "high" (2 steps down from "ultimate" mode) graphics in fullscreen, when I play multiplayer mode. The game runs smoothly and normally for about 10-30mins until the system suddenly freezes on a single frame, locks up with audio stuttering and repeating one annoying short high pitched sound. After that, I can only do a hard reset. Thinking it might be a StarCcraft II problem, I fired up Darksiders and I received the same results.

Things I noticed during the system lockup:
1) There are no anomalies with the frame it freezes up on, no colour distortions nor artifacts.
2) The power is cut from the keyboard and I'm unable to crtlaltdel or attempt to go back into windows + cursor is frozen in spot as well.
3) I set screensaver to 3mins at one time and left the computer alone for 10mins after a lockup. Screensaver does not come on. Nothing else changes as well in that 10mins.
4) Monitor does not lose signal.
5) I opened up the casing and noted that everything inside is still running. The GPU, CPU and other fans are still running as per normal. Mobo LED light shows no error.
6) It definitely happens at around 30mins or before.
7) Of all the solutions I tried, the result is the same and I never get any BSODs nor crash to desktops nor error messages.

I've used the computer for long hours for web surfing, msn, music, worked loooooooong hours on Photoshop and After Effects and I never run into any problems, lags or errors. The system lockup happens exclusively to these 2 games and the lockup definitely happens within 30mins of running them.

With no clue on what is the root of the problem (no error messages!), I blindly searched online for similar issues and tried various solutions.

Possible solutions I've tried:

1) Disable on board sound driver (Realtek HD Audio) from BIOS and unplugged speakers
2) Uninstall and disable NVIDIA High Definition Audio
3) Disable Realtek HD Audio Manager
4) Ran NVIDIA Control Panel: Set to Single Display Performance Mode (from Multiple Display Performance Mode) and set Power Management Mode to Prefer Maximum Performance (from Adaptive)
5) Ran BIOS and set to Asus Optimal (from Normal) for performance and changed back to Normal after testing it didn't help
6) Did clean uninstallations (safe mode + Driver Sweeper) and installations of the following NVIDIA drivers: 270.48, 275.33, 280.26, 285.27(beta)
7) Ran Memtest for hours, no errors
8) Re-insert the RAM cards
9) Ran OCCT GPU test for hours, no errors, system is stable while GPU is running at up to 90 degrees
10) Checked and replugged wires
11) Reinstalled lastest sound drivers/Direct X
12) BIOS updated to the lastest version

Possible solutions I might try:

1) Cleanly format the HDD and reinstall everything again one at a time to test if there was any 1 software/program/driver screwing things up
2) Borrow and try my friend's ATI 5770 to check if its the GPU compatibility issue


Would really appreciate if anyone has suggestions or if you require me to provide more information on something, please let me know, I'll do my best.
September 25, 2011 7:16:07 PM

what are your temps when gaming, or when furmarking? the lack of artifacting makes me suspect CPU temps. Use EVGA precision to track GPU temps, and speedfan to track CPU temps (use the graphs to track temp over time). The key will be trying to stop it just prior to failing to see what the temps are?

Or open the side of the case blow a big fan at it and see if it takes longer than 30mins to lock up.
m
0
l
September 26, 2011 3:47:08 AM

I downloaded the programs to track and here is the screenshot of the temps with just MSN/Chrome/Photoshop(with nth opened) on.

http://i.imgur.com/szMCp.jpg

The CPU temp at 80c freaked me out so I downloaded CPUID HWMonitor to check too and I get the same results. Core temps are hovering around 40c. This is with the side of my case open. I held my hand near the CPU but it doesnt feel like its emitting 80c of heat.

I played StarCraft II for about 15mins and got the following graphs. The Cores temp going up marks the start of the program. (the sudden spike downwards at the end of the graphs is because I closed the game)

http://i.imgur.com/ehDNV.jpg

CPU hovering at 50+, Cores at 60+ and GPU at 60+ too

Also I checked BIOS on system startup and it shows CPU at 50c and mobo at 30c
m
0
l
Related resources
September 26, 2011 6:49:56 PM

Check out this thread:

http://en.community.dell.com/owners-club/alienware/f/37...

A bunch of Alienware Aurora R-3s were having this same problem, including mine. Sudden freezes with no error messages, hard reboots being the only way to recover. After many months it looks like it's faulty RAM that was used. The brand name was Kingston. I know you stated you reseated your RAM but I would advise trying a different RAM brand. Dell replaced the Kingston RAM with Hyundai and it seems to have cleared up the problem.
m
0
l
September 26, 2011 7:15:46 PM

Hi Shich, I'm quite concerned at the CPU@80 in the first image, although not sure why its not represented in the second image.

If 80 is real, then I'd suggest that this is a real issue. Can you confirm that 80 is real?

I seem to recall that there is an issue with C2D/Q's whereby the temp would be inverted, which matches with your first image where the core temp went up whilst the cpu temp went down. its because the temp is measured as a temp gap to max, so if the max is 95 and you are actually at 20 then you have reading equating to 75, if you are actually at 75 then you have a reading equating to 15. putting a -1 into the equation somewhere gives results like yours and I have heard of that before.

Can you try running core temp and see what it says? Also read the C2D temp sticky, it will explain it better.

Having said all that I'm less concerned about temps now, gpu 60 is not an issue, core 60 is a bit of an issue, cpu is wrong.

try core temp, come back. Bionic may have a point.
m
0
l
September 27, 2011 3:27:05 PM

Sorry about the unclear 2nd graph in my prev post, the cpu graph was actually the green line that dipped downwards. I suspect my cpu readings are inverted, when the game starts up the graph dips when it should become warmer.

I ran core temp and speedfan for this test, with my casing covered. Before I start the game:



After the game:



I could try and borrow a friend's ram to try in the next few days.

I really appreciate the help from you guys.
m
0
l
October 1, 2011 1:19:17 AM

I was looking around for SC2 crash problems and I decided to try the popular debug of limiting the framerate. It allowed my game to run up to about 1hr but the same freeze and lock happens.

frameratecap=60
frameratecapGlue=30

I wonder if that is a clue to why my system keeps freezing while playing games?
m
0
l
October 4, 2011 7:47:32 AM

if it is a clue it tells me that we have a gpu issue here, try running at the lowest resolution possible with v-sync on. Idea is to unload the gpu and load up the cpu, if you do not have v-sync on then the gpu will run at 100%, we want it to be at about 10%, and then tell us how long it runs for. and you'll get the nice bonus of it looking really nice when you turn it back up again.

if it runs for a long time like this then we know it could be a gpu cooling issue, although normally they downclock to protect themselves.

If you could borrow a monitor so that you can run speed fan, precision etc. on a second screen it might be really helpful.

winxp used fill up memory 'from the bottom' and so if you had some dodgy memory near the top, it would only be seen when you are using nearly all of your ram, win7 randoming assigns memory usage. I'm not inclined to think it is memory as you 'd be seeing crashes in windows for this reason. But it is worth trying to swap it out.
m
0
l
October 4, 2011 7:54:12 AM

Hang on a sec, what psu do you have, can you run a graph of 12V and other voltages in speedfan, i'm suprised to see the voltage changing on coretemp when you are at the same clocks/idle state.

i'm thinking that the psu is getting warm and stressed and voltage is drooping a little. Which would explain why its sounds temp related giving the timing of the symptoms but not actually having any temp issues being obviously evident.
m
0
l
October 7, 2011 3:17:48 PM

I ran a voltage graph for before:



And after awhile of Darksiders:



Darksiders freezes up in about 15mins, much faster than SC2 does. I run it at my native resolution 1920x1080 and there isn't graphic quality settings.

I tracked voltages for the rest, and similarly they show straight lines.



I ran SC2 in the lowest graphics possible with Vsync on and it ran for 2+hrs without freezing up, I hadn't have time to do a longer run of the game yet.



I ran 3DMark and Memtest (again for a longer time this time)
[3D Mark Basic (free version) results: http://3dmark.com/3dm11/1957885?show_ads=true&page=%2F3...
Memtest ran for 36hrs with no errors, just to be sure.]



One more thing occurred that caused my system to freeze too in the same way (hard resets, no warning signs). I was ripping dvds and when I got to the 3rd dvd, it froze. Other than the dvd ripping software, I have MSN and Chrome on at the time it froze. After I reseted, I ran Coretemp as I attempted to rip the dvd again. Core usage hovered at 95%-100% each and temps rose to 70c. After the ripping got to roughly 70% done, the system froze again.
m
0
l
October 11, 2011 3:37:11 AM

I agree, this sounds like an older issue I had with an nvidia card, actually all the nvidia cards I've had to exactly that when they were overclocked, they froze,2-3 seconds of music or sound then it just keeps repeating the same sound over and over again in loops or one long sound. Sometimes it would kick out of it in about 30 seconds to a minute. I know we aren't talking about overclocking but what your saying sound just like my old nvidia gpu shutdowns. I mean word for word of what you said.
m
0
l
October 11, 2011 8:03:53 AM

OK, so an unloaded GPU and loaded CPU (low res) runs for a long time, (i'm assuming that 2+ hours is not normal and would have crashed).

Can you leave speedfan in the background just tracking 12V, in fact can you run two speedfans, one showing just 12V graph and one showing GPU temp only. Run darksiders for as long as you dare and then come out of it and quickly screen shot the speedfan screens. Better would be to screen shot an evga precision screen so that we can see the GPU fan speeds etc.

I'm not concerned about the CPU temps at this time, they all look good.

I hate to say it but I'm really thinking your PSU is not the best in the world and might be causing your problems.
m
0
l
October 12, 2011 12:55:38 PM

Hmmmmm, I could try down clocking my GPU if it comes down to it. My gpu comes factory overclocked to 900MHz, up from 822MHz core clock.

I ran Darksiders for roughly 14mins. Freezes can happen from 15mins onwards so this is as much as I dared to run it. Here's the screenshot of the graphs of 12V and GPU temp and info from EVGA Precision.



Im afraid I do not have a spare PSU lying around for me to try, but if your suspicions point towards it, I could try asking around.
m
0
l
!