ok: i noticed that the cpu was going to 100x9 (900 Mhz) and i found out later by removing it from the socket that it was really too much.
On the cpu was written 600/128/66 and i think they are respectively the frequence/L2 cache/fsb.
So i manually set the jumpers (don't know if it is the correct word) in the mainboard to have the fsb automatically gained, instead of have it set to 100, as it was.
Now the bios show me that the cpu is running (correctly i hope) at 66x9 (600 Mhz)... but i had the same a few crashes after this change: is there the possibility that some hardware problems occurred in relation to this non-wanted overclock ? The machine is currently up as usual and apparently nothing (very) bad happened, if we don't think about the crashes, so, could the mainboard have been damaged someway ?
I'm sorry but i'm not able to show U the errors i get on screen...this time they're different, at least it seemed so to me, but still fatal, and there is no trace about them in the syslog..maybe it's not able to log anything as soon as the crash occurs.
This is frustrating cause i myself am not able to see what kind of error it is.
Second thing: i swapped the NIC to another PCI slot just half an hour ago.
I'm waiting for something now to happen...hoping nothing will do.
I think it's all i can say now...maybe tomorrow i will report some more, after a few hours of uptime.