[blukatz92]

Okay, here's a good one for your smart people here. I built myself my first PC not too long ago (about two months ago). I have since been experiencing random system crashes during gaming only. It's fine on the internet, running basic programs like Open Office, Windows Media Player and other utility programs. However, if I try to play any games (such as Skyrim, Dirt3, Sonic Generations, etc.) I will get random crashes without warning. It sometimes happens with minutes of starting my game, other times it could take hours. It doesn't matter how much is going on too, sometimes it's when there's a lot of action happening other times it will freeze while I'm just paused or in a menu. It does not crash to my desktop, it just freezes and stays that way until I hard reboot. My mouse and keyboard no longer work (I tried ctrl-alt-del) and if there was any music/sounds playing, this loud, fast clicking noise comes from my speakers. I have repeatedly checked my temps with multiple different programs and I have never had anything get too hot. I've updated all my drivers, updated my BIOS, modified my RAM timings/voltages (the default settings are different from the manufacturer's recommended settings), run AVG and Malwarbytes virus scans, I've run Memtest and passed 6 times, I've moved my GPU to different slots in case the slot itself is bad, I've run Prime95 for 24 hours with no errors. I just don't know what else to try. It just doesn't make sense!

Here's my specs, btw-
Phenom II x4 970BE
ASUS M5A88-V EVO
8GB G.Skill Ripjaw X Series (2x4GB)
HIS Radeon 6850 1GB (I upgraded to this from a Nvidia 9600 GT)
OCZ ZS 650W PSU
Seagate Barracuda 500GB SATA 3Gbps (I upgraded this from an old WD 200GB IDE HDD)

I upgraded the GPU and HDD to not only have newer hardware, but to hopefully fix the problem, but it hasn't made a difference. Any ideas? I'm at a total loss here... :-(

[ElfBane]

Are you OCing? If so, throttle it back to normal and see what happens.

[compnovo]

Hi and welcome to the forums,
Since your temps are OK the symptoms you're describing might be caused by faulty or improperly seated RAM. I suggest removing and reseating the RAM (if you haven't already), and if that doesn't help try running with one stick at a time (4GB is more than enough) --- it's rare to have two bad sticks and this will at least rule out the memory as a problem. BTW, I just had problems with a bad stick of RAM that passed memtest, it isn't foolproof.

[snorg]

Do you get BSOD?
If so does it reference a driver?
If so post its file name here.

[waldojim]

Try running OCCT in cpu linpack mode, then gpu mode, then power supply test mode. When/if occt causes a failure, post those results.

[blukatz92]

@ElfBane- no, I'm not OC'ing, the only modifications to CPU and GPU speed I have are the auto throttle and power saving options in BIOS that my ASUS motherboard came with, and all that does is reduce speed when it's not being used to save power.

@compnovo- I have reseated the RAM multiple times, I've used the primary blue slots for both sticks and I've tried the secondary black slots (ASUS recommends the blue slots for best results, but I tried the other two in case of faulty slots). I have tried using one stick at a time, and I always pass memtest at least 4 times before I'll stop. I'll then try playing a game and every time it freezes again.

@snorg- Actually, as surprising as it may sound, I have been using Windows PCs since the 98 days (given, I was like 7 years old back then) and I have never encountered the BSOD. This is the first major PC issue I have ever had to deal with before.

@waldojim- what is OCCT? Is that a stress tester? I have used Prime95 (when I first built my PC, it was the first test I ran) for 24 hours straight without a fail. I've also used furmark and 3Dmarkvantage to test my system and I always get good results, never fail. It's literally only when I run 3D games that I encounter freezing.

I did come across one detail that I'm unsure of... apparently some AMD CPUs aren't compatible with G.Skill 1600Mhz RAM in duel slot mode, so I tried G.Skill's fix to downclock it to 1333Mhz with tighter timings... no luck. Then I checked the motherboard compatibility, and guess what? My RAM isn't listed. Could that be it?

[waldojim]

blukatz92, on 08 March 2012 - 11:02 PM, said:

@ElfBane- no, I'm not OC'ing, the only modifications to CPU and GPU speed I have are the auto throttle and power saving options in BIOS that my ASUS motherboard came with, and all that does is reduce speed when it's not being used to save power.

@compnovo- I have reseated the RAM multiple times, I've used the primary blue slots for both sticks and I've tried the secondary black slots (ASUS recommends the blue slots for best results, but I tried the other two in case of faulty slots). I have tried using one stick at a time, and I always pass memtest at least 4 times before I'll stop. I'll then try playing a game and every time it freezes again.

@snorg- Actually, as surprising as it may sound, I have been using Windows PCs since the 98 days (given, I was like 7 years old back then) and I have never encountered the BSOD. This is the first major PC issue I have ever had to deal with before.

@waldojim- what is OCCT? Is that a stress tester? I have used Prime95 (when I first built my PC, it was the first test I ran) for 24 hours straight without a fail. I've also used furmark and 3Dmarkvantage to test my system and I always get good results, never fail. It's literally only when I run 3D games that I encounter freezing.

I did come across one detail that I'm unsure of... apparently some AMD CPUs aren't compatible with G.Skill 1600Mhz RAM in duel slot mode, so I tried G.Skill's fix to downclock it to 1333Mhz with tighter timings... no luck. Then I checked the motherboard compatibility, and guess what? My RAM isn't listed. Could that be it?

Prime 95 is a great test to check for stability - with a 1 hour quick pass/fail sort of result. BUT OCCT stresses components in ways that Prime cannot. Grab OCCT and try with it. Run for no less than one hour each to test for stability.

As for your question, I doubt it.
Motherboards get tested with a handful of ram. There are hundreds of different models, and even more variations on those models on the market. 99% of the time, compatibility is not an issue.

This post has been edited by waldojim: 08 March 2012 - 11:13 PM

[blukatz92]

Just downloaded it, although two of the tests have issues based off of the site's description- the GPU Memtest is only for Nvidia cards, I'm running a Radeon, so that one might not work. The power supply test, it says not to use no-name power supplies, because it can fry them. Is my OCZ PSU good enough that I would be safe running this, or should I stay away from it?

[blukatz92]

No, I'm all on stock settings right now.

[coastie65]

blukatz92, on 08 March 2012 - 11:02 PM, said:

I've also used furmark and 3Dmarkvantage to test my system and I always get good results, never fail. It's literally only when I run 3D games that I encounter freezing.

I did come across one detail that I'm unsure of... apparently some AMD CPUs aren't compatible with G.Skill 1600Mhz RAM in duel slot mode, so I tried G.Skill's fix to downclock it to 1333Mhz with tighter timings... no luck. Then I checked the motherboard compatibility, and guess what? My RAM isn't listed. Could that be it?

On the first point, If you are not actually playing in the 3D Mode, then if possible you may want to disable 3D ( I don't know if that is possible with that card ). I can tell you, that when The Witcher 2 was released there was some problems asociated with 3D, but that was with the Nividia card. Most, uninstalled the 3D drivers as they were seperate at the time ( I just simply didn't enable them ).

As to point two, it could be. Try some ram that is listed.

[waldojim]

blukatz92, on 08 March 2012 - 11:31 PM, said:

Just downloaded it, although two of the tests have issues based off of the site's description- the GPU Memtest is only for Nvidia cards, I'm running a Radeon, so that one might not work. The power supply test, it says not to use no-name power supplies, because it can fry them. Is my OCZ PSU good enough that I would be safe running this, or should I stay away from it?

It should not hurt your PSU unless there is something wrong with it. That is actually the point behind part of the test. As for the GPU mem test, don't worry about that part. OCCT has a slightly tweaked copy of Furmark that is NOT throttled by the AMD or Nvidia drivers. SO it will essentially get full run of your video card. The OCCT power supply test is exactly what it sounds like. It places your machine under a complete and total 100% load, and stresses everything. The point is, to make sure your power supply actually remains stable under load. The OCZ should handle this quite easily. However, every company is entitled to one or two bad parts.

[LiveBrianD]

blukatz92, on 08 March 2012 - 11:02 PM, said:

I did come across one detail that I'm unsure of... apparently some AMD CPUs aren't compatible with G.Skill 1600Mhz RAM in duel slot mode, so I tried G.Skill's fix to downclock it to 1333Mhz with tighter timings... no luck. Then I checked the motherboard compatibility, and guess what? My RAM isn't listed. Could that be it?

Most likely, if your CPU isn't compatible with your ram's speed, it'll get underclocked to a compatible one. (ex. I put 1600MHz ram in here but it was underclocked to 1333MHz since that's all the chipset supports unless I overclock) As waldojim said, it's pretty unlikely that ram that's not listed won't work. There are tons of different models of ram out there, and how is the motherboard manufacturer supposed to test all of them? Aside from the branding, many of them are nearly the same! 1333MHz and 1600MHz ram is very common, and is unlikely to be a problem.

[LiveBrianD]

blukatz92, on 08 March 2012 - 11:31 PM, said:

Just downloaded it, although two of the tests have issues based off of the site's description- the GPU Memtest is only for Nvidia cards, I'm running a Radeon, so that one might not work. The power supply test, it says not to use no-name power supplies, because it can fry them. Is my OCZ PSU good enough that I would be safe running this, or should I stay away from it?

OCZ is probably good, being the manufacturer of PC Power & Cooling PSUs after all. The reason it says to not use no-name power supplies is because they typically can't provide their rated wattage for more than a second or two, and thus might blow up if you subject them to that kind of load. Quality power supplies can provide their rated wattage consistently. Also, manufacturers often overrate the PSU wattage you need. In my case, I tried torturing my CPU (prime95) and GPU (OCCT) AT THE SAME TIME. The outervision psu calculator says I need 303W minimum, 353W recommended (100% CPU TDP btw), and as I recall (if I remember correctly), I was drawing 300W from the wall. That's 255W from the components at 85% efficiency (which this PSU is certified for at 50% load).

[blukatz92]

Well, my PSU is rated for 650W, but according to various PSU calculators, my system should be fine with even a 500W, 550W to be safe. OCZ seems like a reputable company, so I went with them. I ran the OCCT CPU test and GPU test and both passed with no errors. I'm running the PSU test as I type this, another 30 minutes left until it finishes. (I'm using my laptop at the moment) So far everything's running fine, the highest temp my CPU has hit so far is 56C, while my GPU has peaked at 70C. No errors yet.

[LiveBrianD]

Those temps are perfectly normal.

[waldojim]

If everything passes, then you are looking at software problems. Try updating drivers - or even changing to older known good drivers for your video card, sound card, network, etc. If this does not fix the problem, then testing with a new install of Windows (don't activate until you prove the problem!) would be the best bet.

[blukatz92]

Unfortunately, it still doesn't tell me why I'm having freezing issues :-P

The power supply test just finished, again it passed. It's one of those things where I'm really glad to see everything is working fine, all these passed tests prove that my system is stable under heavy conditions, but for some reason, if I try to play any of my games, it fails on me. I thought that the stress tests are supposed to push the stability of a PC beyond normal conditions? I mean, so far I've run tests that max out my CPU, GPU and RAM usage and they all pass. But if I try playing a game, it crashes. I've reinstalled every one of them, and to be honest, I really doubt 4 different games would all be having installation errors, especially since each one crashes randomly. I don't even know how to recreate a crash incident on purpose short of playing the game until it does freeze.

[LiveBrianD]

Have you tried reinstalling the video driver?

[blukatz92]

Yeah, I downloaded the latest catalyst driver on AMD's page and reinstalled them. I have heard of a specific method involving a driver sweeper program and using safe mode to install the drivers, but I have not tried that method yet. Does it make any difference, in your opinion?

[LiveBrianD]

My idea: uninstall the video card in device manager, deleting the driver software with it, reboot, reinstall the video driver, reboot again. Then do you get those crashes?