Diagnosing Problems with Computer Memory (RAM)
Memory problems, unless they are catastrophic (-total) failures, often result in unpredictable and seemingly unrelated events on your PC. This is because all your computer's instructions pass through the memory - and it is often pot-luck which instruction lands in the problem RAM addresses.
A hardware problem is classified as a fault in an actual component - or in the configuration of that component - leading to issues with the computer
If you suspect problems with your PC memory, the first thing you need to do is to confirm it really is the memory that is at fault. This can be done in a number of ways:
Firstly, check the Linux log files around the time(s) that the problem occurred to see if there are any clues. The syslog is probably the best one to start with. Check for the string segfault in the logs. Try and match the times that errors show up in the log with your Problem Log to see if they match up: if they often coincide, then they could be connected
Download the (free) Memtest86 tool and burn it's .iso image to CD/DVD or USB memory stick. Reboot your PC using the boot disc or USB and the utility should immediately load and start running. The picture on the left shows the tool running the tests and the one on the right shows a typical error picked up by Memtest86:
The tool should be left running for as longs as possible: at least an hour and possibly longer (-e.g. if your PC problems occur once every four hours of usages, you should consider leaving the tool running for four hours plus). Memtest86 will run continuously until you press the Esc key, to reboot the machine
If you have a spare DIMM available that is compatible with your motherboard (-or the cash to buy one) , try removing the original DIMM(s) and replacing it with the one that you know works: if the problem persists, then it is not likely to be the DIMM itself that is at fault. If the error does go away (-and you had more than one DIMM installed originally) - try different combinations of the DIMMs to narrow down which one is the culprit
All of the above are contributary evidence: one piece on it's own will not secure a conviction but the more evidence you have, the more confident you can be that the problem is memory related!
In general, the process for fixing memory problems is as follows:
If the memory DIMM is not functioning at all (-or several sticks seem to have the same problem), then refer to your motherboard documentation to see if there is some memory setting that needs to be updated in the BIOS setup program for the memory to work correctly
Check the product number of the RAM DIMM(s) installed is listed as compatible on your motherboard manufacturer's website: if not, skip to step 4
Check on your motherboard manufacturer's website if there is a later version of the BIOS that fixes a similar or related problem: if so, consider downloading this and flashing the BIOS. Note: do not interrupt the BIOS update once started as this can result in an unusable system
Finally, purchase a compatible DIMM and follow the RAM Upgrade Process to swap the existing DIMM(s) for the new one