Is there a case where the correct option to cpp is "-openmp"? At least on our local cluster, we use gnu cpp with icc. fpp on Intel doesn't seem to take either option.
On 12/10/2013 02:39 PM, Erik Schnetter wrote:
The CPPFLAGS flags are added to the CFLAGS or CXXFLAGS for compiling. Since the latter probably include -fopenmp or the equivalent, not setting them will still lead to correct code.
However, CPPFLAGS are also used to determine dependencies. Not adding -fopenmp here means that the dependencies will incorrectly exclude anything OpenMP related. In most cases, this will make no difference, but if you have code such as
#ifdef _OPENMP # include "another_header_file.h" #endif
then the dependency on "another_header_file.h" will be missed, potentially leading to wrong code. (I think this would be very unlikely, though.)
-erik
On Dec 10, 2013, at 13:55 , Zach Etienne zachetie@gmail.com wrote:
That's a very good point. I'm curious, does CPP_OPENMP_FLAGS need to be set for any compiler? I've never set it, and have run successfully with OpenMP using pretty much every major compiler out there (albeit with older versions of ET).
On Tue, Dec 10, 2013 at 1:40 PM, Yosef Zlochower yosef@astro.rit.edu wrote: Aren't these options for icc. Does cpp even need -openmp under intel?
On 12/10/2013 01:26 PM, Zach Etienne wrote: At least for my (newish) version of the Intel compilers, the flag "-fopenmp" is equivalent to "-openmp". Might want to try that instead.
-Zach
On Tue, Dec 10, 2013 at 12:57 PM, Steven R. Brandt <sbrandt@cct.lsu.edu mailto:sbrandt@cct.lsu.edu> wrote:
Yes. I hate intel for making their flag "-openmp". Cheers, Steve On 12/10/2013 10:33 AM, Yosef Zlochower wrote: > Hi, > > Every now and then I rediscover a bug that I make when I use > one of the simfactory option files as a starting point for > our local cluster. Since our local cluster is similar to > Lonestar, I start from one of the Lonestar configuration > > Howerver, the option > " CPP_OPENMP_FLAGS = -openmp" > Leads to failed compiles because > cpp -openmp BLAH causes the output of cpp to go to a file > called penmp rather than stdout. > I just wonder if other people were bitten by this. > > On Noether, this caused the compile of GSL to fail > (btw, only GSL, the rest of ET compiled fine). > > _______________________________________________ Users mailing list Users@einsteintoolkit.org <mailto:Users@einsteintoolkit.org> http://lists.einsteintoolkit.org/mailman/listinfo/users
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users
-- Dr. Yosef Zlochower Center for Computational Relativity and Gravitation Associate Professor School of Mathematical Sciences Rochester Institute of Technology 85 Lomb Memorial Drive Rochester, NY 14623
Office:74-2067 Phone: +1 585-475-6103
yosef@astro.rit.edu
CONFIDENTIALITY NOTE: The information transmitted, including attachments, is intended only for the person(s) or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and destroy any copies of this information.
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users