100% RAM usage equals instant power-down?


#1

I have had an issue for a while where the computer will suddenly do an instant power-down. I thought it might be the power supply so that’s been replaced. I have also seen cases where a web page would have a video or other animated picture on it and that would use up enough CPU time that the monitoring program would pop up and warn me that the CPU temperature was climbing past 65°C. However, when this power-down happens, that monitor doesn’t have time to even appear, so it’s not likely to be a case where the computer is shutting down because it’s trying to save itself from overheating.

Normally, there’s no warning that this is going to happen. Today, I had a little bit of warning that something was wrong, but by the time I realized I probably should shut programs down, the power cut out.

At the time it happened, I had one web browser going with over 100 tabs open. Some were for research I was doing and the rest were things I was going to get back to afterwards. I couldn’t get a video to play in it and a couple other pages to load, which will sometimes happen if I’m reaching the limit of what that browser can support in a session, so I switched over to a different browser. That browser started responding more and more sluggishly, and just as I was going to close it, that’s when the computer crashed.

I didn’t have Task Manager open at the time, but I’m wondering if the reason why the computer is crashing is because one of the cores in the CPU spikes at 100% usage and Windows 7 Pro just can’t cope with it. It’s an AMD FX-8350 8-core processor at 4.00 GHz on an Asus M5A99X EVO v1.0 motherboard.


#2

This sounds like overheating to me. Are all of your fans working properly? You might take a look at your settings in the BIOS to see what temperature it shuts down at.


#3

As far as I can tell, all fans are working. I’ll dig into the BIOS settings, but the thing is, the the AI Suite II program that Asus provides for monitoring fan speed and temperature flashes the alert that the temperatures is going up, so I’ll run the full program to see what’s going on. I’ve watched the temp slowly climb from 65° up to as high as 80° in 20 to 40 seconds, and the computer doesn’t crash then. If it was a temperature problem with the CPU, it would have to spike over 15° in less than a second in order for the alert to never appear.


#4

Power supply would have been my first suspect, too.
What if it isn’t the CPU but something on the motherboard that is overheating? Maybe try leaving the side open and a small desk fan blowing at the innards? Of course, if it doesn’t happen fairly frequently, that might not be a helpful troubleshooting step.


#5

I might check the contact material between the CPU and the heat sink. It can dry out over time and lose efficacy. Maybe even a new heat sink.


#6

Likely out of chemlight batteries, or blinker fluid.


#7

Organic, please.


#8

The case is actually open right now on one side. I came home to find the alert waiting for me and it was at 71°C. I don’t have it set to log the readings so I don’t know how long it was like that.


#9

Check CPU heat sink paste… most probably it dried out or doesn’t transfer the heat well away enough.

If it is not that, run memtest86+ for a couple of loops to test for memory errors

Still nothing - run a CPU benchmarking utility and see if the PC bombs out then. If so, then one of your CPU cores may be flakey.


#10

It’s a union CPU. Get the knee breakers in, that’ll straighten it up.


#11

I’ll check the heat sink paste. It’s probably the standard type that AMD includes with the fan, rather than something like Arctic Silver. I just remembered that a couple years ago, I needed to replace the CPU fan for noise or some other reason and the paste had become a cement that refused to let go of the processor. Because I couldn’t get at the release lever, the CPU came out with the fan and several of the pins got left behind in the socket.


#12

Dude… yikes!


#13

It may be the Otter web browser that’s doing it. I watched it last night get more and more sluggish over the course of about three minutes and the temperature alert never popped up. Then, boom. Power instantly cut out again. It looks like it’s not able to cope with heavy demands


#14

Yeah, it’s not temperature. The alert has been popping up for the last half hour as the temperature has been hovering at just over and below 65°, thanks to Corel Paint Shop Pro running in the background. I’ve watched the monitor program report temperatures as high as 80° before, but I was able to see the temperature gradually go up. If this was a thermal shutdown issue, the temperature would have to jump over 20° instantly and it probably would cause another shutdown if I tried rebooting immediately. An immediate reboot has been successful before.

I’ll try stress-testing Otter at some point in the future to see if I can force it to choke and crash without any other programs running.


#15

Topic has been changed from CPU usage to RAM usage.

It wasn’t intentional, but I did a stress-test on the Vivaldi browser this morning. I had a lot of tabs open in about 5 or 6 windows and saw it getting a bit sluggish. Glanced over at Task Manager to see how much memory it was using and saw that the CPU usage graphs for all eight cores were jumping up and down in the 80-100% range.

And just about the time I noticed “Physical Memory: 97%” at the bottom of the window and was seeing what I could shut down to free up memory, another instant power-down.

Maybe the problem is Windows runs out of RAM and just can’t cope any more. Or it’s the combination of high CPU usage and high RAM usage that is causing it to crash. 16GB has worked well for a long time. Maybe it’s time for that old cliche of “Add 4MB and retry” updated to say “Add 16GB and retry”.

Shouldn’t Windows 7 Pro 64-bit be better at memory protection than that?


#16

I’d probably be running memtest


#17

I’m still leaning towards temperature issues. Maybe not with the CPU. Possibly with the video card or the motherboard. Your MB or memory chips could be overheating. I’d install CPUID HWMonitor and keep an eye on things.

Software side… The OS shouldn’t shut things down but you can do some pretty funky things with GPO’s. You could dig through the settings, but I don’t think you’ll find something. It’s a possibility though. Same goes with anything else. You wouldn’t believe (well you might) how often I’ve troubleshoot some issue to find it relates to some obscure service or program running in the background.

Good luck though.


#18

The issue pretty much stopped happening and I never really did find out what was causing it. I think it happened only one more time, but I’ve been hearing intermittent noises coming from the computer case that sounds like a fan or motor issue. Today, one of the hard drives disappeared from the computer which happens to hold most of my data, but it came back up after a reboot.

I’m guessing that maybe the drive is having intermittent problems and it’s sometimes enough to lock up the entire system hard enough to do the power-down.

The file backup is running now. I probably should also clone the drive since it’s got the partition where some of the programs are installed and pagefile.sys.


#19

Compootahs be purdy things. They do weird and wonderful things.

Hope you are on the right track with your hard drives.

Offtopic : A great way to stress test a system properly is to run this :

The free version will be a good way to start things rolling, you can also play around with settings to make your PC work harder or not.

You’ll need a way to monitor your CPU (and other critical components) in case of overheating since it drives your PC really, really hard. Use http://www.cpuid.com/softwares/hwmonitor.html for that.

I’ve used this combination to stress-test three or four refurbished laptops (all done by myself) and they’re still running.


#20

The insta-crash has happened only once or twice in the last year. It’s probably the video card doing it. Changing the power settings to not turn off the monitor after a while has helped, and I make sure not to leave Firefox open too long if I don’t have to, especially if it’s on a website that auto-refreshes or has videos on it. The CPU temp warning is more likely to go off when videos are running through Firefox.

Not too worried about it right now. I probably will do an OS reinstall during my vacation next month. There’s a couple of things on this Win 7 computer that aren’t working quite right, such as System Restore somehow got turned off a while back.