Re: Seg 11 in GCC
In message <[🔎] 65lIw6ljcsB@khms.westfalen.de>, Kai Henningsen writes:
>mdorman@lot49.med.miami.edu (Michael Alan Dorman) wrote on 29.03.96 in <m0u2f>Hx-00029GC@lot49.med.miami.edu>:
>> So, I've not changed the hardware, and I'm excercising it more than I
>> was previously (keeping the load above 6, mostly), and yet I see no
>> SIG11s, even during the parallel compilations. That would tend to
>> cast significant doubt on the common assertion that SIG11 = hardware
>> problem, no?
>You have *one* case where the reason *may* be something different.
How else can you explain the fact that I tried _several_ times to
compile a single file (conmakehash.c), and got SIG11s until I
reinstalled GCC, at which point it worked? Did my hardware magically
get better? Or, could it be that it was a corrupt executable?
And if it was, in this instance, a corrupt executable, how can you
assert that doesn't cast doubt on the commonly-held wisdom that
_every_ SIG11 is a RAM error?
I'm not saying that RAM's not suspect---I would be a fool to try and
pretend that I know more than Bruce, who obviously has the smarts to
play with this stuff at a much deeper hardware level than I---just
that I think the assertion that it's _always_ RAM is wrong.
>On the other hand, there are at least thousands of cases where it
>*clearly* was a hardware problem - hardware fiddling made it go away.
>Apply the math yourself.
Oh, come now, unless you can document your assertion of "at least
thousands of cases where it *clearly* was a hardware problem",
"Applying the math" hardly matters because your data is suspect.
As far as your obvious disagreement with my last paragraph, I think
it's because I didn't make the point strongly enough in my first
paragraph that I don't doubt some SIG11 issues are RAM related but
that I object to
>the current trend of automatically classifing every single SIG11 as
>indicative of bad memory
Would you have had as much problem with my last paragraph if I had
said, "That would tend to cast significant doubt on the common
assertion that a SIG11 in GCC can _always_ be inextricably linked to a
hardware problem, no?"
Mike.
--
"Don't let me make you unhappy by failing to be contrary enough...."
Reply to: