Bizarre difference in uptimes for same kernel

Rob Partington (rjp@faethor.wamphyri.net)
Tue, 9 Sep 1997 21:55:45 +0000 (GMT)


A little while back, when 2.1.51 was the bleeding edge, it was
downloaded in my normal "more haste, less speed" thirst for the bleeding
edge. Configured it, compiled it, rebooted into it, used it. Thought
"ok, this is fine", installed it for a colleague.

One week later, my machine has rebooted about 90 times[*], his once.
I'm a tad puzzled by this, and go through all the available hardware in
order to check that it's not some rogue hardware driver (ncr scsi vs
adaptec aic7, eepro100 vs clone ne2000), only constant hardware was the
sound card which my colleague also has. Oh, and the motherboard and
processor are exactly the same (P166, SuperMicro HX), but I have 64M, he
only 48M.

So, as you can imagine, I'm a bit puzzled why my average uptime is about
40 minutes, and his is 3 days, especially given that his machine does an
equal amount of work, if not more (I'm the sole user of this machine, he
has two people "leeching" X off him.)

The only difference I can find is that he's using pgcc to compile his
kernel, and I'm using gcc. So I grab the kernel off his machine, stick
it on mine, and my uptime (before rebooting into 2.1.54) was 2 days.

Is it possible that compiling the kernel with pgcc instead of gcc
actually makes for a more stable kernel? I don't know enough about the
kernel, or about gcc/pgcc to judge, so I throw it open to the massed
wisdom of linux-kernel for an answer...

Any information appreciated...

[*] solid lockups, no oops, no console switching, no SysRq

-- 
Rob Partington, rjp@wamphyri.net
Running Linux since, well, quite a long time after everyone else, it
 looks like...

-- 
Rob Partington, rjp@wamphyri.net, rjp@netlink.co.uk
Senior Systems Chocolate