4 tests failed on gmp 5.0.3 using mingw64 gcc 4.7

[Rene Sugar]

On 2/9/2012 4:50 AM, Torbjorn Granlund wrote:
> Rene Sugar<rene.sugar at gmail.com>  writes:
>
>    $ config.guess
>    k10-pc-mingw32
>
>    gcc says k10 is not a valid value. I tried modifying config.status to compile with
>    amdfam10 instead of k8 and got the same 4 errors.
>
> I don't understand what you mean.
>
> If you want us to change anything with the configure system, you need to
> clearly specify the behaviour you think is incorrect.  See also the GMP
> manual, gmplib.org/manual/Reporting-Bugs.html.
>
> GMP seems to correctly identify your CPU as k10.  If you pass "k10" to
> gcc somehow (how?), isn't that a usage issue, or perhaps a gcc issue?
>

I was saying that changing the code generation by altering the -mtune 
and -march did not affect (improve) the test results.

http://gcc.gnu.org/onlinedocs/gcc/i386-and-x86_002d64-Options.html

Makefile:CFLAGS = -O2 -pedantic -m64 -std=gnu99 -mtune=k8 -march=k8
Makefile:CXXFLAGS = -O2 -pedantic -m64 -std=gnu99 -mtune=k8 -march=k8


It does not appear to be a compiler problem as you originally suggested.


The 4 crashes were fixed by applying the following patch:

On 2/8/2012 4:55 AM, Marc Glisse wrote:
 > On Tue, 7 Feb 2012, Rene Sugar wrote:
 >
 >> I got the following 4 errors on gmp 5.0.3 when running "make check" 
on a mingw64 build on Windows 7 64-bit.
 >
 > Hello,
 >
 > could you check if the patch I posted there:
 > http://gmplib.org/list-archives/gmp-bugs/2012-January/002524.html
 > helps and post your results on the mailing-list? It seems possibly 
related.
 >
 > (I am not sure the patch applies cleanly to 5.0.3, but I hope it 
does. Otherwise you may need to test a snapshot.)
 >