Some games freezing, can sometimes alt+tab out, usually requires hard shutdown

Toggle sidebar Toggle sidebar
D

dmr83457

Reputable
Mar 24, 2014
6
0
4,510
ASUS M5A78L-M/USB3 (firmware up to date)
AMD FX-6350
Corsair Vengence 2x4GB 2133MHZ
EVGA GTX 660 SC 2GB (drivers up to date)
Western Digital WD10EZEX 1TB
Inland 500W ATX ILG-5900R2

While playing some games the screen will freeze and last sound will be in a loop. Essentially just the game is frozen and I can alt+tab to other applications and move the mouse around. Windows explorer and other apps are often frozen but not always. If I have game in window when it freezes and I can alt+tab to all windows but they might not respond. Sometimes if I leave the computer up with frozen game (CSGO) it will drop back to lobby screen saying I timed out.

I put this build together about 2 months ago and initially only had an issue with Serious Sam 3. I noticed that they had issues with some AMD chips in the past so just chalked it up to some unpatched issue. However I played Mirror's Edge a lot and eventually started having freezes occasionally there. Now I am big into Counter-Strike: Global Offensive (CSGO) and getting freezes very often. Sometimes I can't go more than 5 minutes without freezing while other times I can go an hour or two. Since competitive mode penalizes abandoning a game I end up restarting and joining game again in the hopes it won't freeze again which is hit or miss.

I have run memtest86, OCCT, Furmarker and Windows/WesterDigital/Seagate disk diagnostics and not finding the problem. A few weeks ago I got a freeze during GPU test on OCCT but have not been able to replicate the issue. I have kept ResourceMonitor, OpenHardwareMonitor and OCCT open while in CSGO to try to see what is causing the freeze but haven't gained any really useful info, but I did notice that disk usage was a steady 10Mb/s.

Any suggestions on what to check out? Any definitive check I can run to rule out each component as a problem?

TESTING

Memory - Memtest86: Previously I ran the standard test with no errors. This time I ran the single pass test with no errors found.

CPU - Prime95: Small FFT run for 2 hours, saw vcore fluctuations from ~1.44v to .888v

PSU - OCCT: Ran power supply test for a while, was working in another room and noticed that OHM web server was no longer accessible. System had restarted and was on AMI BIOS screen that does not normally show up on boot as ASUS has their own boot screen that shows up instead. It ended with "SATA3G_1 Hard Disk:S.M.A.R.T. Command Failed ... Press F1 to Resume". I pressed F1 which did nothing so I hit ctrl+al+del, the system restarted and came up with nearly the same screen. I flipped PSU off for maybe 10 seconds, flipped back on, then started the system up without an issue.

Disk - WinDLG (Western Digital Data LifeGuard): Ran extended test which it passed

HARDWARE CHANGES

PSU - Corsair CX600M - I caught an in-store sale and rebate on this unit at local MicroCenter last night. I installed it and so far no freezing.
 
Solution
Onus
I would have suggested memory first, but it sounds like you've tested that pretty thoroughly. Next up is the PSU. I'm not familiar with the brand. If it is actually a PSU-shaped object, it might be good for only a fraction of its label, possibly not cleanly. Particularly since your problems seem to happen under load, it may be getting stressed and wandering out-of-spec, causing crashes.
Another possibility, if your last remark indicates that your hard drive is constantly active, is to rebuild your swap file in case it has become corrupt.
Sort by date Sort by votes
Onus

Onus

Titan
Jan 27, 2006
37,037
12
89,965
I would have suggested memory first, but it sounds like you've tested that pretty thoroughly. Next up is the PSU. I'm not familiar with the brand. If it is actually a PSU-shaped object, it might be good for only a fraction of its label, possibly not cleanly. Particularly since your problems seem to happen under load, it may be getting stressed and wandering out-of-spec, causing crashes.
Another possibility, if your last remark indicates that your hard drive is constantly active, is to rebuild your swap file in case it has become corrupt.
 
Upvote 0 Downvote
Solution
D

dmr83457

Reputable
Mar 24, 2014
6
0
4,510


I am running diagnostics and stress tests again and will document as I go along. I get the feeling though that the symptoms such as game freeze but lack of entire system locking up or black screen would indicate a specific issue.
 
Upvote 0 Downvote
D

dmr83457

Reputable
Mar 24, 2014
6
0
4,510
Just had voltage fluctuations on motherboard vcore value. It is usually right around 1.344v but has had some dips down to 0.888v. It would seem that is not kosher. Correct?
 
Upvote 0 Downvote
D

dmr83457

Reputable
Mar 24, 2014
6
0
4,510
I think the vcore fluctuations during Prime95 stress test and reboot during OCCT Power Supply and inability to startup until PSU off/on indicates a PSU problem. Anyone have any other specific tests to run before I try another power supply tonight?
 
Upvote 0 Downvote
D

dmr83457

Reputable
Mar 24, 2014
6
0
4,510


I don't have another PSU on hand so may test the current one further or more directly. I have a multimeter and looking up specific instructions for testing with it but not sure if a "stress" test is possible. However might not be worth going through the trouble as this was not an expensive PSU (obviously) so if I have to just get a good new one and don't return the current one then no big loss. If doesn't solve issue then can return one of them. Will update with more info later. Thanks
 
Upvote 0 Downvote
D

dmr83457

Reputable
Mar 24, 2014
6
0
4,510


Ok, I got a new PSU (update in quesion) that has done the trick so far. I was able to play CSGO for 2-3 hours last night with no problem. I have some additional questions that you may be able to answer

1. After startup vCore, reported by PC Probe 2, fluctuated between 0.9 and 1.3 but after starting the game it stayed right around 1.3-1.4v. I have high performance enabled in windows but have not made any other adjustments. Is this caused by vdroop which I have been reading about and something I should adjust in the motherboard to improve consistency in performance?

2. The GTX 660 has a 6 pin power plug. When installing the new PSU I noticed it had 2 plugs and cables labelled for PCI-E. When I built my computer a couple months ago I actually used the Y-cable that came with the graphics card which splits the 6 pin plug into two 4 pin plugs which I then connected to the connectors for the old PSU. After removing the old PSU though I noticed that it does have a single 6+2 cable that would have worked for the graphics card. Is it possible that the problems I experienced are related to the way I wired up the graphics card, causing part of the PSU to get overloaded unnecessarily?
 
Upvote 0 Downvote
Onus

Onus

Titan
Jan 27, 2006
37,037
12
89,965
vCore will vary based on load, as a part of the power saving features. Under heavy load, it will stay high; nothing looks unusual there.
It is unlikely that your method of wiring made any difference. If you overload a +12V rail that has an OCP circuit on it (any quality multi-rail PSU), the PSU will shut down, not respond as yours did. I think your entire PSU was overloading, and it wandered out of spec and caused crashes. Hopefully it did no damage to your other parts.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom