Stability Checks

[greggerm 312]

Simmers,Over the past 24 hours, I've noticed some unfortunate stability issues with my PC. At first I though it might have been viral or malicious software that snuck through my net, but that has since been ruled out. I was getting some Windows desktop anomalies in that Firefox needed to be minimized then restored in order to be used, icons couldn't be dragged to the recycle bin or to/from Windows Explorer, and even a crash of my normally rock-stable email program. This morning, I restored my PC from a virgin-fresh TrueImage build I created back when I last rebuilt my computer - no applications on it, and nothing to cause any incompatibilities. Some of the symptoms were still there... until suddenly they "snapped" free and the computer has been stable ever since. This instability this morning and the snap to stability has all been observed using the computer via a Windows Remote Desktop connection.... not directly at the desktop. I've ruled out viral and malicious software as the cause, as my morning rebuild would have purged all bad stuff out. I have ruled out CPU overheating, as this morning core temperatures were in the 40's. I am now running Orthos on the machine (and have been for an hour and a half) and core temps are in the low to mid 50's with no hiccups so far. I will attempt to run MemTest when I get home from work, but does anyone have any other diagnostic tests which would be applicable to "general system instability"?Power supply seems like it's good, as I am fully loading the CPU and hard drives with stuff to do and nothing has hiccuped this afternoon. The video card has given me issues in recent past with some screen blanking and veeery slight white noise but only on hellaciously hot days, indicating a potential overheat. Checking it's temps I'm showing normal for the card, and the instability I am experiencing comes quickly after powering the computer on, so the card hasn't had a chance to heatsoak. I've got a ton of components now wish-listed at Newegg, as this might be a great excuse to update this PC and move it's components elsewhere, but I'd like to get more use out of them. Your commentary will be appreciated!-Greg

[Guest D17S]

Just start swapping things out. I just had a fatal problem. Intermittent no-boot to windows with any Nvidia driver installed. Safe mode with Vista's VGA driver was the only reliable mode. 3 new vcards didn't help. Ended up being the 2nd (unused, ATM) vcard had fragged and was taking down the whole system, on occasion. Logic has its place, but with this stuff, it's often more trial and error (with lots of the later).

[greggerm 312]

Thanks Sam,Given the computer seems to be pretty stable now that the video card and it's drivers aren't processing my work (as I am using a remote desktop session to test from work), AND given that I had some sporadic issues with it in recent past, it may be time to swap her out. I know you're not specifically recommending video card swaps (as it just happens to be what was wrong with yours), but if my MemTest runs clean tonight (and my Orthos processor test runs clean this afternoon), that will be the next major component to play with. Amazingly, Best Buy and Newegg have the same price on a 4870 ($285), so that might be my first upgrade/test. If it works, great! If it's still broken, I've updated my gfx card in preparation for my next major update. -Greg

[greggerm 312]

ISSUE RESOLVED:Would you believe a bogus USB-PS2 mouse adapter caused the problems? (or the PS/2 mouse plug on the motherboard went south) In prepping to put in the new video card, I noticed my mouse cable was under a lot of strain. I re-seated the cable and nothing really changed. Since it is a USB mouse plugged into a PS/2 port via an adapter, I decided to just go native and plug it into a USB port. Badda-bing, badda-boom - all my problems disappear. The mouse works 100%, no drag-and-drop issues, no browser issues... nothing. Just goes to show you that even the seemingly insignificant components can contribute to larger problems.Now I have an unopened 4870 to return to Best Buy. I couldn't use it anyway, as my power supply upon review is only a 450W with a single PCIe lead. Thank goodness it stayed sealed in it's box! =--------------------------------------------------=On my graphics card side, my BFG 8800GT OC was running VERY hot, as mentioned above. The fan seemed to be locked at 29%, not adjusting for any loads. The stability issues I was having seemed independent of video, but I did some research just in case. It turns out that BFG issued a BIOS update in 12/2007 for the card which "unlocked" the fan, allowing it to speed up and slow down with temperature changes. I flashed the card and now things work properly. Using a GPU Stress Test program, I kicked the card in the pants hardcore and the card spools up to 85 degrees C (nVidia's reference sweetspot temperature under heavy load) and the fan ramps up in speed to match it. Watching closely in SpeedFan and RivaTuner, the GPU will get to 85 degrees and the fan will rapidly spool up to 100%, and they stay there. This indicates the temperature sensors, fan, and card are working in synergy... (100% load, 100% fan, spot-on reference temperature) I've operated the card WITHOUT this fan fix for some time - I can only imagine that the card got upwards of 100 degrees (design max is 110 before it starts shutting down).Interestingly, FSX does not come anywhere near fully loading the GPU, even with high AA and AS settings. When I did some manual configuration of the fan speeds, I had the fan running 100% while flying FSX in clouds, and the GPU was down at 64 degrees!!! 85 degrees at 100% under load, 64 degrees at 100% under FSX. Further proof (as if we needed any more) that GPUs aren't terribly stressed when flying MSFS. =--------------------------------------------------=Well, I hope this helps someone down the road, which is why I followed up my initial post.