Green rectangles appear, system is immediately unresponsive and then reboots

[matt_symes]

Hi

Hmm. You said that when you switched to the console it did not crash after a couple of minutes extra; enough time to get the logs.

Does it crash if you leave it in the console for a long period of time ?

I'm just trying to find some pattern here.

Kind regards

[Palu]

While wandering around menus, I ran some sort of rescue console and took a picture because I saw a complaint about ACPI. I don't know what it means, but I'll post it here and do the bare bones strip you mentioned.

http://www.danwolf.net/images/IMG_0074.JPG

[Palu]

Also, it looks like it'll never crash in the console.

[matt_symes]

Hi

Also, it looks like it'll never crash in the console.

Interesting.

Just another simple test to perform.

When it locks up, if you press the caps lock key, does the caps lock led change ?

Kind regards

[Palu]

I started to open my computer and realized that there is really nothing to disconnect in your instructions besides the mouse and one memory stick. The wifi and bluetooth are built in to the motherboard. There are no card readers, CD drives, secondary HDs, or other devices plugged in. The graphics (HD 4000) is integrated from the processor too. That said, I'm at 7 minutes since startup without a crash. Usual time to crash is 30 seconds to 4 minutes, I'd say, so... maybe the stick of RAM I removed was bad. Does that explain not crashing in the console with bad RAM though?

[Palu]

Another thought: Looking up ACPI, I saw that it's about power management, and I remembered that my tiny case has a 300 watt power supply. Since I have no CD drive, an SSD, and a third gen i5, I figured that would be enough, but is it possible that the memory stick I removed is fine but it pushes my machine just over the edge in terms of power? How would I test that?

[matt_symes]

Hi

As far as i can see we have a problem with the mtrr registers and the memory addresses they hold.

We also have resource conflicts as highlighted in your last photo and one of my previous posts.

I am really beginning to think this is due to the mtrr registers but that is a guess.

This really does look like some kind of memory addressing issue.

We can set values for that using kernel boot parameters and i think we should look at doing that and see where that takes us.

I'm still looking and thinking. This is not an issue i have come across before.

EDIT:

I figured that would be enough, but is it possible that the memory stick I removed is fine but it pushes my machine just over the edge in terms of power? How would I test that?

Try it without that stick for a while but memory is not a power hog.

Kind regards

[Palu]

3 hours stable... if I have to use 8 GB, it's a shame, but really that's plenty. I'm not a gamer but a programmer, so I think most my strength will be from having my first SSD. I'll peek back periodically to see if you have anything else you'd like to check.

[matt_symes]

Hi

@palu.

Are you still getting the mtrr problems with half the ram ? Check your logs and post back.

If they are still there i would suspect power supply and power to drive the exra dimm.

If they are not there i would suspect those mtrr problems.

Kind regards

[Palu]

My computer ran fine all night and into today. Looked at the hours since I noted stability, and kern.log showed no more "mtrr_cleanup: can not find optimal value" or the memory error it was listing yesterday morning and afternoon (well, I suppose it was your evening).