Bill Allombert on Sat, 09 Oct 2004 19:36:42 +0200


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: machine freezes


On Sat, Oct 09, 2004 at 11:41:48AM +0200, Dieter Hink wrote:
> Hello,
> 
> I am using
> 
> AMD Athlon 64 Processor 3200+
> Linux SuSE 9.1 32-Bitmodus
> gp 2.2.7 32 Bit-Version compiled with gcc-3.3.3 (SuSE Linux)
> 
> This little script (also in attachment) looks for social numbers, beginning
> with 276, here starting after iteration 823. After a short while, gp says
> "MPQS will take several hours", then the system freezes. Same effect with
> SuSE 9.1 64 Bitmodus and gp 2.2.7 64-Bit-Version.
> 
> I do extended Hardware-Test with memtest for 24 hours or like repeated
> Kernel-Makes like in www.bitwizard.nl/sig11 for 30 hours. After test
> anything looks good.

I cannot reproduce (on my very old testbox).

What happen is that MPQS try to factor
704429103247132813360069420564100477852723109104419894453196792185824193732364461

which will take some time with PARI implementation.

If you are adventurous, try to do
\g9
factorint(704429103247132813360069420564100477852723109104419894453196792185824193732364461,6)

you can try to save the output by issuing
\l/path/some/file
before factorint,
but be sure /path/some/file resides on a device with no important data.
A very fast device would be best like a flash memory device.

As for the freeze cause: this code do basically 2 things: (unsigned)
multiplication/division and heavy I.O. Given that I would blame
the kernel, as a random guess.

Cheers,
Bill.