Thinkpads Forum

I've got a C2D T60 with the GMA950 graphics chip. I've been getting these random stuck pixels which turn back to normal if I switch the application. For example, I may be playing a movie in Windows Media Player, and I get a white pixel in the black area; if I Alt-Tab to a different application (e.g. Firefox) the pixel goes back to normal (in that case, the page Firefox was displaying used a black background so the pixel reverted to black), and if I switch again to WMP the pixel turns white. I have come to the conclusion that a faulty RAM module is causing this. Has anyone experienced the same problem? Would you say I'm right in suspecting the memory (since system memory is being used as video RAM)? Thanks.

download memtest86+ and start a memory test. You can also try 3D Mark, it will stress your graphic card and his memory.

madrat wrote:I've got a C2D T60 with the GMA950 graphics chip. I've been getting these random stuck pixels which turn back to normal if I switch the application. For example, I may be playing a movie in Windows Media Player, and I get a white pixel in the black area; if I Alt-Tab to a different application (e.g. Firefox) the pixel goes back to normal (in that case, the page Firefox was displaying used a black background so the pixel reverted to black), and if I switch again to WMP the pixel turns white. I have come to the conclusion that a faulty RAM module is causing this. Has anyone experienced the same problem? Would you say I'm right in suspecting the memory (since system memory is being used as video RAM)? Thanks.

I really doubt it is a RAM issue. It's probably a stuck pixel. Switching to firefox isn't a very good test for stuck pixels. There are actually utilities for this sort of thing. Try google to find one.

There is very slight chance you could solve the problem with a video driver update (or if you recently updated it could be caused by the update.)

agee with post #2 that memtest will rule out a RAM issue

I don't think the LCD is at fault here. The stuck pixels usually disappear when switching between applications, but they may stay if Windows uses the same video buffer for the output of several applications. The location of the stuck pixels (when they appear) is random as well. I can also make them disappear by switching the screen res to 800x600 and then back to 1024x768 (thereby forcing a change of the physical address of the memory used by the video buffer), or by playing a 3D game.

So definitely not a funky LCD. Everything points to the memory being the culprit, but I thought I'd check if anyone else in the forums has had the same problem anyway.

I tried memtest86 once before, and it didn't find any errors. I might have run it with caching enabled, though. I'll try it again with all caches disabled, letting it run all night, as well.

random locations for stuck pixels (or any pixel anomoly) would tend to be video memory or related hardware..

also could be a bad driver..

but it is most unlikely to really be a stuck pixel..
though these CAN come and go but always from the same location..

I've ruled out defective graphics hardware and video drivers, since in either case the problem wouldn't stop at a single stuck pixel, but be more widespread (e.g. line streaks and crashes).

I can't explain why the memory would get stuck, though. And why the stuck part would span no more than 4 bytes (the max size a screen pixel occupies in memory).

Another thing I'm going to try is remove one of the two RAM modules and then look for stuck pixels with remaining module running. Then swap if I don't see any stuck pixels, and observe again.

madrat wrote:I've ruled out defective graphics hardware and video drivers, since in either case the problem wouldn't stop at a single stuck pixel, but be more widespread (e.g. line streaks and crashes).

I can't explain why the memory would get stuck, though. And why the stuck part would span no more than 4 bytes (the max size a screen pixel occupies in memory).

Another thing I'm going to try is remove one of the two RAM modules and then look for stuck pixels with remaining module running. Then swap if I don't see any stuck pixels, and observe again.

Not trying to be argumentative, just trying to help BUT...

You can't r/o hardware or driver by the above statement.

You can't r/o stuck pixels unless you run a utility that rules it out. Again, you just can't say well, its not because...

memtest will r/o the RAM as a possibility. Sure it doesn't check all of the RAM, but you can test it once. Swap the slots that the RAM are in and run memtest again. That will pretty much test all of the RAM.