Jason Grout wrote:
> Dr. David Kirkby wrote:
>> I'd be interested what you get if you build this program, which was
>> written by one of the gcc guys to try to get to the bottom of this issue
>> with mpfr not building.
>>
>> On the Sun T5240 ('t2') donated by Sun to the Sage project, it dumps core:
>>
>> kir...@t2:[~] $ ./a.out
>> n=0
>> n=1
>> Abort (core dumped)
>>
>> On my own Blade 2000 it runs like this:
>>
>> drkir...@kestrel:[~] $ ./a.out
>> n=0
>> n=1
>> n=2
>> n=3
>> n=4
>> n=5
>> n=6
>> n=7
>> n=8
>> n=9
>> n=10
>>
>> What does it do on your machine?
>
>
> It runs like on your Blade 2000.
>
>
> CPU: Mobile Intel(R) Pentium(R) 4 - M CPU 2.00GHz
>
> OS: Ubuntu 9.04
> GCC: gcc (Ubuntu 4.3.3-5ubuntu4) 4.3.3
>
> $ ./test_program
> n=0
> n=1
> n=2
> n=3
> n=4
> n=5
> n=6
> n=7
> n=8
> n=9
> n=10
>
>
> Jason
Thank you.
This looks like it is a bug in the Sun implementation of memset() of
Solaris in the library used on for the Sun T2+ proessors.
i and j are both 2^31, so their sum is 2^32, which should be 0 in 32-bit
code (which this is). So the program should replace 0 bytes of memory
with the number 0 (in other words, do nothing.)
But clearly it is doing something as it corrupts the value of the buffer.
This should be a relief to the developers of gcc, mpir and mpfr!! Sun
was the last party I expected to be guilty, but that is where I think
the fault lies now.
Dave
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
