>>>>> "guillaume" == Guillaume Cottenceau <[EMAIL PROTECTED]> writes:

guillaume> Paolo Pedroni <[EMAIL PROTECTED]> writes:
>> Il 12:43, marted� 30 ottobre 2001, hai scritto:
>> 
>> > Do you use it often? Did it allow you to detect memory problems in the
>> > real world?
>> 
>> I detected memory problems at least three times, using memtest. It just works!
>> Now, everytime some friend of mine asks me to check their malfunctioning 
>> computer, first thing I do is stick a memtest floppy disk in their drive and 
>> check their memory.

guillaume> Personally, I use for a long time the simple following thing: I recompile
guillaume> 100 times a kernel, storing the logs, and I then verify all the logs are
guillaume> the same ; when memory or chipset or processor are malfunctioning,
guillaume> sometimes GCC receives signal-11 because of failing hardware.

guillaume> I'm wondering if "memtest" would not miss some of the errors; also it
guillaume> doesn't use the harddisk so it can miss chipset problems related to disk
guillaume> probably.

guillaume> Of course, "memtest" is really more easy to use than recompiling a kernel.
memtest is better for testing memory.  But recompiling the kernel is
perhaps better for finding that you have overclocking/bad CPU fan &
similars.

There is another program, cpuburn, that help to found that kind of
problems, it executes secuences of instructions that are known to heat
the CPU.  I don't remind the URL, but goople should help you.

Later, Juan.


-- 
In theory, practice and theory are the same, but in practice they 
are different -- Larry McVoy

Reply via email to