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

Re: SIN/COS broken in gfortran on mipsel? Test request

Hello GCC and mips* folks,

Kevin B. McCarty wrote:

> I find it hard to believe, but it seems that the SIN() and COS()
> built-ins might be broken in gfortran on mipsel.

Here's an update on this, in case anyone was wondering.

Thiemo very kindly gave me an account on a mips machine, where the same
problem with gfortran-4.3 occurs as on mipsel.  The upshot is that
gfortran-4.3 tries to optimize successive calls to SIN() and COS() by
generating object code that makes a single call to sincosf() in libm.
Somehow the format of that function call is not working.

It seems to be a gfortran bug rather than a libm bug, since equivalent C
code that explicitly calls sincosf() gives correct results.  I submitted
the bug upstream at http://gcc.gnu.org/PR35662 for anyone else who wants
to follow it.

best regards,

Kevin B. McCarty <kmccarty@gmail.com>
WWW: http://www.starplot.org/
WWW: http://people.debian.org/~kmccarty/
GPG: public key ID 4F83C751

Attachment: signature.asc
Description: OpenPGP digital signature

Reply to: