On Monday January 27 2003 11:08 am, Praedor Tempus Atrebates wrote: > Some of what I have found with regards to messages like: > > Unable to handle kernel paging request at virtual address 01100112 > > is that it is thought to be related to memory/cache problems. Does > anyone know one way or another? I have been running memtest86 for > over an hour on the problem system and so far there have been no > errors.
memtest86 is very lightweight. Even it's name is misleading as there is no software test for just memory. When you run a diag on a system, the cpu/cache (both L1 and L2)/motherboard chipsets/ram all have to work together to produce no errors. A better test is, ftp://mersenne.org/gimps/mprime2212.tar.gz Quick start: unpack to a dir, look for the executable mprime*. run it with ./mprime, it'll ask if you want to join the prime search or just test, opt for test. Then run it with ./mprime -m and choose number 17, the torture test. If it encounters any hardware errors, it'll exit an tell you so. BTW, if you're curious about mprime, http://www.mersenne.org/prime.htm The acid test, http://users.ev1.net/~redelm/ While it's billed as a cpu test, as I alluded to above, it stresses the whole system. If you can run cpuburn's modules appropriate for your system or an hour, your system is bulletproof IMO, cpuburn is the quickest way to determine if you need to start lookin at OS (kernel) and software for the root of your problem. But, you should see if you can run mprime's torture test for an hour or two first. -- Tom Brinkman Corpus Christi, Texas
Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com