- SJ On 5/20/2012 7:32 AM, Daniel Russel wrote:
You can add CHOLMOD="no" to your config to turn it off. On May 20, 2012, at 2:16 AM, Yannick Spill<> wrote:Cholmod is optional for isd, and also only deprecated stuff uses it. I'd force the config to compile without cholmod if I knew how. On 20 mai 2012, at 07:22, Seung Joong Kim<> wrote:There is only one libchomod in my mac /opt/local/lib/libcholmod.a Here is the result of "otool -L". Not sure how much helpful this information is.. By the way, I need to enable the "ISD" module, since it is required by the "SAXS_merge" module... If there is no easy solution, I will stick to use the linux machine. Thanks a lot! [sjkim@MAC /opt/local/lib]$ otool -L libcholmod.a Archive : libcholmod.a libcholmod.a(cholmod_aat.o): libcholmod.a(cholmod_add.o): libcholmod.a(cholmod_band.o): libcholmod.a(cholmod_change_factor.o): libcholmod.a(cholmod_common.o): libcholmod.a(cholmod_complex.o): libcholmod.a(cholmod_copy.o): libcholmod.a(cholmod_dense.o): libcholmod.a(cholmod_error.o): libcholmod.a(cholmod_factor.o): libcholmod.a(cholmod_memory.o): libcholmod.a(cholmod_sparse.o): libcholmod.a(cholmod_transpose.o): libcholmod.a(cholmod_triplet.o): libcholmod.a(cholmod_check.o): libcholmod.a(cholmod_read.o): libcholmod.a(cholmod_write.o): libcholmod.a(cholmod_amd.o): libcholmod.a(cholmod_analyze.o): libcholmod.a(cholmod_colamd.o): libcholmod.a(cholmod_etree.o): libcholmod.a(cholmod_factorize.o): libcholmod.a(cholmod_postorder.o): libcholmod.a(cholmod_rcond.o): libcholmod.a(cholmod_resymbol.o): libcholmod.a(cholmod_rowcolcounts.o): libcholmod.a(cholmod_rowfac.o): libcholmod.a(cholmod_solve.o): libcholmod.a(cholmod_spsolve.o): libcholmod.a(cholmod_drop.o): libcholmod.a(cholmod_horzcat.o): libcholmod.a(cholmod_norm.o): libcholmod.a(cholmod_scale.o): libcholmod.a(cholmod_sdmult.o): libcholmod.a(cholmod_ssmult.o): libcholmod.a(cholmod_submatrix.o): libcholmod.a(cholmod_vertcat.o): libcholmod.a(cholmod_symmetry.o): libcholmod.a(cholmod_rowadd.o): libcholmod.a(cholmod_rowdel.o): libcholmod.a(cholmod_updown.o): libcholmod.a(cholmod_super_numeric.o): libcholmod.a(cholmod_super_solve.o): libcholmod.a(cholmod_super_symbolic.o): libcholmod.a(cholmod_ccolamd.o): libcholmod.a(cholmod_csymamd.o): libcholmod.a(cholmod_metis.o): libcholmod.a(cholmod_nesdis.o): libcholmod.a(cholmod_camd.o): libcholmod.a(cholmod_l_aat.o): libcholmod.a(cholmod_l_add.o): libcholmod.a(cholmod_l_band.o): libcholmod.a(cholmod_l_change_factor.o): libcholmod.a(cholmod_l_common.o): libcholmod.a(cholmod_l_complex.o): libcholmod.a(cholmod_l_copy.o): libcholmod.a(cholmod_l_dense.o): libcholmod.a(cholmod_l_error.o): libcholmod.a(cholmod_l_factor.o): libcholmod.a(cholmod_l_memory.o): libcholmod.a(cholmod_l_sparse.o): libcholmod.a(cholmod_l_transpose.o): libcholmod.a(cholmod_l_triplet.o): libcholmod.a(cholmod_l_check.o): libcholmod.a(cholmod_l_read.o): libcholmod.a(cholmod_l_write.o): libcholmod.a(cholmod_l_amd.o): libcholmod.a(cholmod_l_analyze.o): libcholmod.a(cholmod_l_colamd.o): libcholmod.a(cholmod_l_etree.o): libcholmod.a(cholmod_l_factorize.o): libcholmod.a(cholmod_l_postorder.o): libcholmod.a(cholmod_l_rcond.o): libcholmod.a(cholmod_l_resymbol.o): libcholmod.a(cholmod_l_rowcolcounts.o): libcholmod.a(cholmod_l_rowfac.o): libcholmod.a(cholmod_l_solve.o): libcholmod.a(cholmod_l_spsolve.o): libcholmod.a(cholmod_l_drop.o): libcholmod.a(cholmod_l_horzcat.o): libcholmod.a(cholmod_l_norm.o): libcholmod.a(cholmod_l_scale.o): libcholmod.a(cholmod_l_sdmult.o): libcholmod.a(cholmod_l_ssmult.o): libcholmod.a(cholmod_l_submatrix.o): libcholmod.a(cholmod_l_vertcat.o): libcholmod.a(cholmod_l_symmetry.o): libcholmod.a(cholmod_l_rowadd.o): libcholmod.a(cholmod_l_rowdel.o): libcholmod.a(cholmod_l_updown.o): libcholmod.a(cholmod_l_super_numeric.o): libcholmod.a(cholmod_l_super_solve.o): libcholmod.a(cholmod_l_super_symbolic.o): libcholmod.a(cholmod_l_ccolamd.o): libcholmod.a(cholmod_l_csymamd.o): libcholmod.a(cholmod_l_metis.o): libcholmod.a(cholmod_l_nesdis.o): libcholmod.a(cholmod_l_camd.o): On 5/19/2012 8:22 PM, Daniel Russel wrote:I doubt things have ever been looked into for doing static compilation of isd with cholmod. One needs to figure out which libraries cholmod depends on and add them to the add_dependency call (when doing dynamic linking they are brought in automatically). To help us get started, you can do ldd on the cholmod dynamic lib (somehting like /usr/lib/libcholmod.so) and send the results (if you are in a mac, you need to run "otool -L " on the .dylib instead). That shows what libraries the dynamic lib brings in on your platform. If you don't need isd, just disbable it as I'm not sure we want to get this working unless really needed as it could be painful. On May 19, 2012, at 7:41 PM, Seung Joong Kim wrote:Hi Yannick, Suddenly I have a problem in compiling the ISD module as follows. Do you have any idea what happens? I recently upgraded to OSX 10.7 Lion, hopefully that is not an issue.. Thanks a lot! Compiling shared build/src/imp_isd_config.cpp Compiling modules/isd/bin/benchmark_gaussianprocess.cpp Install file: "build/swig/IMP_isd.i" as "build/data/swig/IMP_isd.i" Running swig on file build/swig/IMP_isd.i /opt/local/bin/swig -castmode -interface _IMP_isd -DPySwigIterator=IMP_ISD_PySwigIterator -DSwigPyIterator=IMP_ISD_SwigPyIterator -python -c++ -naturalvar -fvirtual -Wextra -o /Users/sjkim/imp/trunk/build/src/IMP_isd_wrap.cpp-in -oh /Users/sjkim/imp/trunk/build/src/IMP_isd_wrap.h-in -Ibuild/swig -I/opt/local/include -I/usr/include -I/usr/local/include -I/Users/sjkim/imp/trunk/build/include -I/Users/sjkim/imp/trunk -DIMP_SWIG /Users/sjkim/imp/trunk/build/swig/IMP_isd.i /Users/sjkim/imp/trunk/build/include/IMP/isd/bivariate_functions.h:129: Warning 314: 'lambda' is a python keyword, renaming to '_lambda' /Users/sjkim/imp/trunk/build/include/IMP/isd/bivariate_functions.h:129: Warning 314: 'lambda' is a python keyword, renaming to '_lambda' /Users/sjkim/imp/trunk/build/include/IMP/isd/bivariate_functions.h:129: Warning 314: 'lambda' is a python keyword, renaming to '_lambda' /Users/sjkim/imp/trunk/build/include/IMP/isd/bivariate_functions.h:129: Warning 314: 'lambda' is a python keyword, renaming to '_lambda' /Users/sjkim/imp/trunk/build/include/IMP/isd/bivariate_functions.h:129: Warning 314: 'lambda' is a python keyword, renaming to '_lambda' /Users/sjkim/imp/trunk/build/include/IMP/isd/bivariate_functions.h:129: Warning 314: 'lambda' is a python keyword, renaming to '_lambda' /Users/sjkim/imp/trunk/build/include/IMP/isd/bivariate_functions.h:129: Warning 314: 'lambda' is a python keyword, renaming to '_lambda' /Users/sjkim/imp/trunk/build/include/IMP/isd/bivariate_functions.h:129: Warning 314: 'lambda' is a python keyword, renaming to '_lambda' Install file: "build/src/IMP.isd.py" as "build/lib/IMP/isd/__init__.py" Patching swig file build/src/IMP_isd_wrap.h Patching swig file build/src/IMP_isd_wrap.cpp Compiling shared build/src/IMP_isd_wrap.cpp Linking Shared Library build/lib/libimp_isd.dylib Undefined symbols for architecture x86_64: "_dpotrf_", referenced from: _r_cholmod_super_numeric in libcholmod.a(cholmod_super_numeric.o) "_zpotrf_", referenced from: _c_cholmod_super_numeric in libcholmod.a(cholmod_super_numeric.o) _z_cholmod_super_numeric in libcholmod.a(cholmod_super_numeric.o) ld: symbol(s) not found for architecture x86_64 collect2: ld returned 1 exit status scons: *** [build/lib/libimp_isd.dylib] Error 1 scons: building terminated because of errors. - SJ -- Seung Joong Kim, Ph.D. Postdoctoral Scholar, Andrej Sali group Department of Bioengineering and Therapeutic Sciences University of California at San Francisco (UCSF) 1-217-649-2147_______________________________________________ IMP-dev mailing list https://salilab.org/mailman/listinfo/imp-dev-- Seung Joong Kim, Ph.D. Postdoctoral Scholar, Andrej Sali group Department of Bioengineering and Therapeutic Sciences University of California at San Francisco (UCSF) 1-217-649-2147 _______________________________________________ IMP-dev mailing list https://salilab.org/mailman/listinfo/imp-dev_______________________________________________ IMP-dev mailing list https://salilab.org/mailman/listinfo/imp-dev_______________________________________________ IMP-dev mailing list https://salilab.org/mailman/listinfo/imp-dev
-- Seung Joong Kim, Ph.D. Postdoctoral Scholar, Andrej Sali group Department of Bioengineering and Therapeutic Sciences University of California at San Francisco (UCSF) 1-217-649-2147