[Gcc-cfarm-users] gcc64 upgraded to OpenBSD 5.0

Zimmermann Paul Paul.Zimmermann at loria.fr
Mon Jan 30 10:07:16 CET 2012


       Vincent,

> Date: Mon, 30 Jan 2012 03:37:04 +0100
> From: Vincent Lefevre <vincent at vinc17.net>
> 
> Hi,
> 
> On 2012-01-29 20:22:16 +0100, Matthieu Herrb wrote:
> > In order to provide a more recent and useful OpenBSD/sparc64
> > environment, I've upgraded gcc64 to OpenBSD 5.0. 
> > http://gna.org/support/?2866
> 
> There were no problems to build GNU MPFR 3.1.0-p4 on OpenBSD 4.6.
> But after the upgrade, 6 tests are failing on gcc64.
> 
> -- 
> Vincent Lefèvre <vincent at vinc17.net> - Web: <http://www.vinc17.net/>
> 100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
> Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

does GMP "make check" pass? I just tried compiling GMP 5.0.3 and "make check"
fails:

PASS: t-div
*******************************************************************************
mpn_dcpi1_bdiv_qr inconsistent in test 296
N=   00000000001fffff ffffffffffffffff ffffffffffffffff ffffffffffffffff ... 0000000000000000 0000000000000000 0000000000000000 0000000000000000
D=   0001ffffffffffff ffffffffffffffff ffffffffffffffff ffffffffffffffff ... 0000000000000000 0000000000000000 0000000000000000 0000000000000001
Q=   f2a76aa9c3f11342 b4be0a1954168d14 285ca08d115eb3fe 70c149b027bbd015 ... 0000000000000000 0000000000000000 0000000000000000 0000000000000000
R=   fffe1ab12acc781d d97a9683ebcd57d2 e5d7af46bee5dd42 98031e7d6c9fb088 ... c9fed6e1394ba081 8a18b73ced84936b 6e5de2d3f0d86912 9e95639dea78da66
Rb=  1, Cy=1
T=   00000000001fffff ffffffffffffffff ffffffffffffffff ffffffffffffffff ... 0000000000000000 0000000000000000 0000000000000000 0000000000000000
nn = 1401, dn = 174, qn = 1227
*******************************************************************************
Abort trap (core dumped) 
FAIL: t-bdiv

This is with the default GCC version (4.2.1). I'll now try with egcc (4.2.4).

Note that this error is mentioned on the GMP status page:
http://gmplib.org/devel/testmachines/check/failure/loulou.esiee.fr:64

but this issue is not mentioned on http://gmplib.org/#STATUS, where there is
an extra closing parenthesis in "strike this bug.)"

Paul



More information about the gmp-bugs mailing list