On Wed, 2002-12-04 at 03:33, Alex Bennee wrote: > On Tue, 2002-12-03 at 19:27, James Sparenberg wrote: > > This is one point I do have to give to M$ they do seem to be able to > > dynamically map around bad blocks on HDD's and ram a lot better than > > Linux or FreeBSD. > > I don't know if Windows actively maps round bad memory (and in my case > its only a single bit). However it could be the difference in > distribution of kernel-mode structures. If windows keeps them all in one > place but Linux keeps them all over memory it could make Linux more > susceptible. After all corrupt memory in user-mode is never a problem - > you just lose the one process. It would of had to been a kernel > structure getting clobbered to cause the hang, I'm guessing a pageing > request or something like that rather than a syscall. > > > As for the badmem tools I've never been able to get > > them running nor have I really taken a hard run at it. (I cheated and > > swapped memory with my wifes windows box) > > Can't get the memtest tools to compile under Mandrake, failing to link > to the maths library for some reason (and the make needs to be run as > root which is bad).
Can't remember if this is true but if memtest is c++ then I've heard rumor that it is causing problems with some of the older programs that haven't been updated to gcc3.02's differences. Witness the fact that mozilla is compiled under 2.96 for MDK. That could be the trouble. James > > > The other thing I have seen > > is that sometimes running mem at a slower speed works as well. (bad bit > > at 133 but not at 100). As for DDR memory... haven't got any so don't > > know. > > After spending a good hour on the phone to PC world who siad they can't > diagnose the problem unless I'm running windows I swapped the sticks. > Hopefully this should solve the problem. My college will find out how > well Windows will cope soon. > > > > > James > > > > On Tue, 2002-12-03 at 08:50, Alex Bennee wrote: > > > On Tue, 2002-12-03 at 13:05, Alex Bennee wrote: > > > > <snip> > > > > bit can cause the machine to lock up completely and mess with > > > > X windows Ctrl-Alt-BS. I'll re-run memtest tonight and see if the > > > > results are consistent in failing. > > > > > > I know i've got a memory problem but the Ctrl-Alt-Backspace is not > > > related. Its seems to be a power save feature (it works even in Grub) > > > but Windows seems to disable it by the time it boots. I've tried both > > > noapic and apm=off kernel options but I still cannot disable it. I can't > > > help feeling the two may be related, although I did seem to be able to > > > kill X without killing the machine once its hardly consistent. > > > > > > -- > > > Alex Bennee > > > Senior Hacker, Braddahead Ltd > > > The above is probably my personal opinion and may not be that of my > > > employer > > > > > > > > > ---- > > > > > > > > Want to buy your Pack or Services from MandrakeSoft? > > > Go to http://www.mandrakestore.com > > > > > > > > ---- > > > > > Want to buy your Pack or Services from MandrakeSoft? > > Go to http://www.mandrakestore.com > -- > Alex Bennee > Senior Hacker, Braddahead Ltd > The above is probably my personal opinion and may not be that of my > employer > > > ---- > > Want to buy your Pack or Services from MandrakeSoft? > Go to http://www.mandrakestore.com
Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
