Troubleshooting, debugging memory problems

There're a number of "tool compilatons" that the maintainers package into a bootable CD image (ISO) that you can also write to an USB drive. Here're the best tool sets that I know of:
  • Hiren's BootCD (AFAIK this one is not downloadable from the maintainer's page since the ISO contains a number of not perfectly legal progs)
  • Ultimate Boot CD
  • UBCD4Win (Ultimate Boot CD for Windows)
The latter (UBCD4Win) is built around BartPE (Bart's Preinstalled Environment), thus it's not an ISO, but a builder that will create an ISO for you (and requires an WinXP or Win2003 install CD during the process).

Each of these toolsets contain utilities for testing CPU+RAM operation. However you should be aware that malfunctioning does sometimes occur only under extreme load. Thus memory testers like memtest86+ will not trigger them since they test only a part of the memory at a time. You need some stress testing tools for the job. Eg. among the Windows tools in UBCD4Win you'll find Prime95 (P95) which is used by many for stress testing CPU+memory. In the original Ultimate Boot CD you'll find similiar tools too, eg. the Mersenne Prime Test.

I already had PCs in my hands that worked normally under regular load, but during high load weird things (crashes, file system corruption, process deaths) started to happen. Memtest86+ did not produce any errors, but using memtester and stress I could determine that the memory modules were the problem (and it turned out they were overheating).

Comments

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

stressapptest

Google released a stress testing package named "stressapptest".

Stressful Application Test (or stressapptest) tries to maximize randomized traffic to memory from processor and I/O, with the intent of creating a realistic high load situation.
Stressapptest may be used for various purposes:
  • Stress test: as described here.
  • Hardware qualification and debugging.
  • Memory interface test: see the Theory behind this.
  • Disk testing.

Might be worth checking out.

Syndicate content