问题描述
我正在尝试使用f2py
编译将openMP
用于python的fortran例程.
I am trying to compile a fortran routine that uses openMP
for python using f2py
.
这是文件bsp.f90
:
module OTmod
!$ use omp_lib
implicit none
public :: get_threads
contains
function get_threads() result(nt)
integer :: nt
nt = 0
!$ nt = omp_get_max_threads()
!$omp parallel num_threads(nt)
write( *, * ) 'hello world!'
!$omp end parallel
end function get_threads
end module OTmod
如果我用它编译
f2py -m testmod --fcompiler=gfortran --f90flags='-fopenmp' -lgomp -c bsp.f90
编译有效,但是将其导入python失败并显示错误
compilation works, but importing it to python fails with the error
ImportError: dlopen(/Users/USER/omp_py/testmod/testmod.cpython-36m-darwin.so, 2): Symbol not found: _GOMP_parallel
Referenced from: /Users/USER/omp_py/testmod/testmod.cpython-36m-darwin.so
Expected in: flat namespace
in /Users/USER/omp_py/testmod/testmod.cpython-36m-darwin.so
但是,如果我只删除构成并行循环的两行
However if I remove only the two lines that make the parallel loop
!$omp parallel num_threads(nt)
和
!$omp end parallel
然后导入到python并执行它可以工作,因此omp_get_max_threads()
可以工作,但是并行循环导致它失败.这是Anaconda和我的brew install gcc
之间的不匹配吗?我该如何解决?
then importing to python and executing it works, so omp_get_max_threads()
works, but the parallel loop causes it to fail. Is this some mismatch between Anaconda and my brew install gcc
? How can I fix this?
下面的系统设置信息
macOS 10.13
macOS 10.13
Python 3.6.2 |Anaconda custom (x86_64)| (default, Sep 21 2017, 18:29:43)
[GCC 4.2.1 Compatible Clang 4.0.1 (tags/RELEASE_401/final)] on darwin
gfortran -v
Using built-in specs.
COLLECT_GCC=gfortran
COLLECT_LTO_WRAPPER=/usr/local/Cellar/gcc/7.2.0/libexec/gcc/x86_64-apple-darwin17.0.0/7.2.0/lto-wrapper
Target: x86_64-apple-darwin17.0.0
Configured with: ../configure --build=x86_64-apple-darwin17.0.0 --prefix=/usr/local/Cellar/gcc/7.2.0 --libdir=/usr/local/Cellar/gcc/7.2.0/lib/gcc/7 --enable-languages=c,c++,objc,obj-c++,fortran --program-suffix=-7 --with-gmp=/usr/local/opt/gmp --with-mpfr=/usr/local/opt/mpfr --with-mpc=/usr/local/opt/libmpc --with-isl=/usr/local/opt/isl --with-system-zlib --enable-checking=release --with-pkgversion='Homebrew GCC 7.2.0' --with-bugurl=https://github.com/Homebrew/homebrew-core/issues --disable-nls
Thread model: posix
gcc version 7.2.0 (Homebrew GCC 7.2.0)
f2py -c --help-fcompiler
Gnu95FCompiler instance properties:
archiver = ['/usr/local/bin/gfortran', '-cr']
compile_switch = '-c'
compiler_f77 = ['/usr/local/bin/gfortran', '-Wall', '-g', '-ffixed-
form', '-fno-second-underscore', '-m64', '-fPIC', '-O3', '
-funroll-loops']
compiler_f90 = ['/usr/local/bin/gfortran', '-Wall', '-g', '-fno-second-
underscore', '-m64', '-fPIC', '-O3', '-funroll-loops']
compiler_fix = ['/usr/local/bin/gfortran', '-Wall', '-g', '-ffixed-
form', '-fno-second-underscore', '-Wall', '-g', '-fno-
second-underscore', '-m64', '-fPIC', '-O3', '-funroll-
loops']
libraries = ['gfortran']
library_dirs = ['/usr/local/Cellar/gcc/7.2.0/lib/gcc/7/gcc/x86_64-apple
-darwin17.0.0/7.2.0']
linker_exe = ['/usr/local/bin/gfortran', '-Wall', '-Wall']
linker_so = ['/usr/local/bin/gfortran', '-Wall', '-g', '-m64', '-
Wall', '-g', '-undefined', 'dynamic_lookup', '-bundle']
object_switch = '-o '
ranlib = ['/usr/local/bin/gfortran']
version = LooseVersion ('7.2.0')
version_cmd = ['/usr/local/bin/gfortran', '-dumpversion']
Fortran compilers found:
--fcompiler=gnu95 GNU Fortran 95 compiler (7.2.0)
otool -L testmod.cpython-36m-darwin.so
/Users/USER/omp_py/testmod/testmod.cpython-36m-darwin.so:
@rpath/libgomp.1.dylib (compatibility version 2.0.0, current version 2.0.0)
@rpath/libgfortran.3.dylib (compatibility version 4.0.0, current version 4.0.0)
/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1252.0.0)
/usr/local/lib/gcc/7/libgcc_s.1.dylib (compatibility version 1.0.0, current version 1.0.0)
@rpath/libquadmath.0.dylib (compatibility version 1.0.0, current version 1.0.0)
推荐答案
我不确定如何解决此问题以进行动态链接.它主要与与LD_LIBRARY_PATH相关的问题有关.
I am not sure how to solve this issue for dynamic linking. It mostly has to do with something related to LD_LIBRARY_PATH or something.
但是,作为解决方法,您可以static
链接库
But as a workaround you can do static
linking of the library
将-static
选项添加到setup.py
extra_link_args=['-lgomp -static']
这将消除在运行时定位依赖项的需求
This will remove the need for locating the dependency at run-time
这篇关于f2py:在fortran中使用openMP并行失败的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持!