Martin Mokrejs added the comment:

I think it would be of tremendous help if python reported the physical memory 
address though do not know how much work is that for you. I already said that, 
I wonder why memtest86+ could not find an error, I even tried other tools for 
that. Could --with-pydebug immediately after writing those pad bytes try to 
read them back, just to make sure they are readable? Theoretically one could 
catch the wrong bit of memory immediately.

It is random, I think CPU must be fully loaded for a long while, plus something 
unknown has to kick in. Maybe that has to heat up the memory modules to cause 
the failure. My motherboard is already 3rd but doubt it is at fault, except 
that the SandyBridge chip is bad by design. Was replaced few days ago with CPU 
cooler (was the main reason for technician to come). I take that replacing 
motherboard was a nice offer from Dell. I would instead think that maybe CPU is 
heating too much.

BIOS is weird. ACPI and PCIe ASPM are exploiting Linux kernel hotplug support 
and PCI developers are puzzled. I experience that SandyBridge resets e.g. its 
USB2.0 controller if there was a hotplug even on some of its other root ports 
(express card was inserted/ejected). But in overall I don't think it is a bad 
deal and it works. These exceptional glitches are I think a software error or 
BIOS error. It just wouldn't be running all the days and nights. No, it is not 
overheated, at least not now with the replaced cooler.

I can put in other memory module (with which I bought the laptop from Dell). 
Other than that, hmm ... I could take out the hard drive and attach it to some 
other computer, boot from Live-DVD ... the problem is I need a lot of RAM for 
some of my stuff and it would have to be relatively same CPU type (I compiled 
binaries with avx). Yes, the emerge crash definitely was another case, I could 
hope to get that happen on even 1GB RAM machine, sure. I just don't have a 
testcase for that to trigger. I tried to stitch something but no, I probably 
coded something not going in the right direction. If you would have the time to 
stitch down some test for me to execute with the garbage collector calls, it 
would be probably the best.

----------

_______________________________________
Python tracker <rep...@bugs.python.org>
<http://bugs.python.org/issue18843>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe: 
http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

Reply via email to