OCCA undefined references

[dcoxe93]

Very new to nekRS so I am having difficulties here. Is there an exhaustive list of 3rd party utilities we need to install?

here is the nature of the error which comes at the end of building nekrs.

I run ./nrsconfig:
-- Found OpenMP_C: -fopenmp (found version "4.5")
-- Found OpenMP_CXX: -fopenmp (found version "4.5")
-- Found OpenMP_Fortran: -fopenmp (found version "4.5")
fatal: not a git repository (or any of the parent directories): .git
-- Using CMake version 3.25.0
-- System : Linux-5.15.0-52-generic
-- Build type : RelWithDebInfo
-- C flags : -Wall -Wextra -Wunused-function -Wunused-variable -Wwrite-strings -Wfloat-equal -Wcast-align -Wlogical-op -Wshadow -Wno-c++11-long-long -fopenmp -O2 -g -march=native -mtune=native -ftree-vectorize
-- CXX flags : -Wall -Wextra -Wunused-function -Wunused-variable -Wwrite-strings -Wfloat-equal -Wcast-align -Wlogical-op -Wshadow -Wno-unused-parameter -fopenmp -fno-strict-aliasing -O2 -g -march=native -mtune=native -ftree-vectorize
-- Found OpenMP_C: -fopenmp (found version "4.5")
-- Found OpenMP_CXX: -fopenmp (found version "4.5")
-- Found OpenMP_Fortran: -fopenmp (found version "4.5")
-- OpenMP include dirs:
-- OpenMP libraries: /usr/lib/gcc/x86_64-linux-gnu/9/libgomp.so
-- Found CUDA: /usr/local/cuda (found version "11.8")
-- Found CUDA: /usr/local/cuda (found version "11.8")
-- Could NOT find HIP (missing: HIP_ROOT_DIR HIP_INCLUDE_DIRS HIP_LIBRARIES HIP_RUNTIME_DEFINE HIP_HIPCONFIG_EXECUTABLE HIP_PLATFORM HIP_COMPILER)
-- MPI include dirs:
-- MPI libraries:
-- MPI link flags:
-- MPI executable: /usr/bin/mpiexec
-- MPI numproc flag: -n
-- Found MPI_UNDERLYING_COMPILER: /usr/bin/g++

----------------- Summary -----------------
Installation directory: /home/daniel/.local/nekrs
C compiler: /usr/bin/mpicc
C++ compiler: /usr/bin/mpic++
Fortran compiler: /usr/bin/mpif77
Default backend : CUDA
CPU backend compiler: g++ (flags: -O3 -g -march=native -mtune=native)
NVIDIA CUDA backend enabled (flags: -O3 --fmad=true -lineinfo)
GPU aware MPI support enabled

-- Configuring done
-- Generating done
-- Build files have been written to: /home/daniel/nekRS-22.0/build

--> Now just run cmake --build ./build --target install -j8

and running cmake to build nekrs I get the following errors:
/usr/bin/ld: libnekrs.so: undefined reference to occa::device::setup(occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyTo(occa::memory, long, long, long, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::memory occa::device::malloc<void>(long, occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyTo(void*, long, long, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::~properties()' /usr/bin/ld: libnekrs.so: undefined reference to occa::kernelArg::kernelArg(double)'

I can't seem to figure out why I have these issues. I have
MPI 3.3.2
CUDA 11.8
gcc 9.4.0
ubuntu 20.04LTS

Thanks,
Daniel

[stgeke]

Are there no other error messages during the build? It looks like occa cannot be build successfully.

[dcoxe93]

Technically the full list of compilation errors are:

/usr/bin/ld: libnekrs.so: undefined reference to occa::device::setup(occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyTo(occa::memory, long, long, long, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::memory occa::device::malloc<void>(long, occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyTo(void*, long, long, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::~properties()' /usr/bin/ld: libnekrs.so: undefined reference to occa::kernelArg::kernelArg(double)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::device::buildKernel(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, occa::properties const&) const' /usr/bin/ld: libnekrs.so: undefined reference to occa::properties::properties()'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::operator=(occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyFrom(occa::memory, long, long, long, occa::properties const&)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::properties(char const*)' /usr/bin/ld: libnekrs.so: undefined reference to occa::device::createStream(occa::properties const&)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyFrom(void const*, long, long, occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::device::buildKernelFromString(std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&, std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::memory occa::device::malloc<void>(long, void const*, occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory occa::device::malloc(long, occa::memory, occa::properties const&)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::kernelArg::kernelArg(int)' /usr/bin/ld: libnekrs.so: undefined reference to occa::properties::properties(std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::properties(occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::kernelArg::kernelArg(float)'
collect2: error: ld returned 1 exit status
make[2]: *** [CMakeFiles/nekrs-bin.dir/build.make:102: nekrs] Error 1
make[1]: *** [CMakeFiles/Makefile2:240: CMakeFiles/nekrs-bin.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
/usr/bin/ld: libnekrs.so: undefined reference to occa::device::setup(occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyTo(occa::memory, long, long, long, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::memory occa::device::malloc<void>(long, occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyTo(void*, long, long, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::~properties()' /usr/bin/ld: libnekrs.so: undefined reference to occa::kernelArg::kernelArg(double)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::device::buildKernel(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, occa::properties const&) const' /usr/bin/ld: libnekrs.so: undefined reference to occa::properties::properties()'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::operator=(occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyFrom(occa::memory, long, long, long, occa::properties const&)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::properties(char const*)' /usr/bin/ld: libnekrs.so: undefined reference to occa::device::createStream(occa::properties const&)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyFrom(void const*, long, long, occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::device::buildKernelFromString(std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&, std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::memory occa::device::malloc<void>(long, void const*, occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory occa::device::malloc(long, occa::memory, occa::properties const&)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::kernelArg::kernelArg(int)' /usr/bin/ld: libnekrs.so: undefined reference to occa::properties::properties(std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&)'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::properties(occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::kernelArg::kernelArg(float)'
collect2: error: ld returned 1 exit status
/usr/bin/ld: libnekrs.so: undefined reference to occa::device::setup(occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyTo(occa::memory, long, long, long, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::memory occa::device::malloc<void>(long, occa::properties const&)' /usr/bin/ld: libnekrs.so: undefined reference to occa::memory::copyTo(void*, long, long, occa::properties const&) const'
/usr/bin/ld: libnekrs.so: undefined reference to occa::properties::~properties()' //usrusr//binbin//ldld:: libnekrs.solibnekrs.so:: undefinedundefined referencereference toto ``occaocca::::devicekernelArg::::setupkernelArg((occadouble:):'properties /constusr&/)bin'/ ld/:usr /libnekrs.sobin:/ ldundefined: referencelibnekrs.so :to undefined occareference: :todevice ::occabuildKernel:(:stdmemory::::__cxx11copyTo:(:occabasic_string:<:charmemory,, stdlong:,: char_traitslong<,char >long,, stdocca::::allocatorproperties< charconst>& )> constconst'& ,/ usrstd/:bin:/__cxx11ld::: basic_stringlibnekrs.so<:char ,undefined stdreference: :tochar_traits <charocca>:,: memorystd :occa::allocator:devicechar:: malloc>< voidconst>&(,long ,occa :occa::properties: propertiesconst &const)& )const''

//usrusr//binbin//ldld:: libnekrs.solibnekrs.so:: undefinedundefined referencereference toto ``occaocca::::memoryproperties::::copyToproperties((void)',
/longusr,/ binlong/,ld :occa :libnekrs.so::properties undefinedconst &reference) toconst ' occa/:usr:/propertiesbin:/:ldoperator:= (libnekrs.soocca:: :undefinedproperties referenceconst &to) '
occa/:usr:/propertiesbin:/:ld~:properties (libnekrs.so):'
undefined/ usrreference/ binto/ ld`:occa :libnekrs.so::memory :undefined: copyFromreference( occato: :`memoryocca,: :longkernelArg,: :longkernelArg,( doublelong),'
occa/:usr:/propertiesbin /constld&:) 'libnekrs.so
:/ usrundefined/ binreference/ ldto: `libnekrs.soocca:: :undefineddevicemake[2]: *** [CMakeFiles/fdm-bin.dir/build.make:102: nekrs-bench-fdm] Error 1
:reference: buildKernelto( std`:occa::__cxx11::properties::basic_string:<propertieschar(,char stdconst::)char_traits'<
char/>usr,/ binstd/:ld::allocator libnekrs.sochar: undefined> referenceconst &to, `stdocca::::__cxx11device::::basic_stringcreateStream<(charocca,: :stdproperties: :constchar_traits&<)char'>
,/ usrstd/:bin:/allocatorld<:char >libnekrs.so :> undefinedconst &reference, toocca :`:occaproperties: :constmemory&:): copyFromconst('void
/constusr*/,binmake[1]: *** [CMakeFiles/Makefile2:321: CMakeFiles/fdm-bin.dir/all] Error 2
/longld,: longlibnekrs.so,: occaundefined: :referenceproperties toconst &`)occa':
:/propertiesusr:/:binproperties/(ld):'
libnekrs.so/:usr /undefinedbin /referenceld :to libnekrs.so`:occa :undefined: devicereference: :tobuildKernelFromString (`stdocca::::__cxx11properties::::basic_stringoperator<=char(,occa :std::properties: char_traitsconst<&char)>',
/stdusr:/:binallocator/ldchar: libnekrs.so>: constundefined& ,reference stdto: :`__cxx11occa::::basic_stringmemory<:char:,copyFrom (stdocca::::char_traitsmemory<,char >long,, stdlong:,: allocatorlong<,char >occa :>: propertiesconst &const,& )occa':
:/propertiesusr /constbin&/)ld :const 'libnekrs.so
:/ usrundefined/ binreference/ ldto: `libnekrs.soocca:: :undefinedproperties :reference: propertiesto( char` occaconst:*:)memory'
occa/:usr:/devicebin:/:ldmalloc:< void>libnekrs.so(: longundefined, referencevoid toconst *`,occa :occa::device::properties: createStreamconst(&occa):':
properties/ usrconst/&bin)/'ld
:/ usrlibnekrs.so/:bin /undefinedld :reference libnekrs.so: undefined referenceto to` occa`:occa::memory: memoryocca::::copyFromdevice(:void: mallocconst<void,> (longlong,, longocca,: :occamemory:,: propertiesocca :const:&properties) 'const
&/)usr'/
bin//usrld/:bin /libnekrs.sold:: undefinedlibnekrs.so :reference undefinedto reference` occato: :`deviceocca::::buildKernelFromStringkernelArg(:std::kernelArg:(__cxx11int:):'basic_string
</charusr,/ binstd/:ld::char_traits libnekrs.sochar: ,undefined stdreference: :toallocator <`charocca>: :>properties :const:&properties,( stdstd::::__cxx11__cxx11::::basic_stringbasic_string<<charchar,, stdstd::::char_traitschar_traits<>,, stdstd::::allocatorallocator<> >> constconst&&),'
occa/:usr:/propertiesbin /constld&:) libnekrs.soconst:'
undefined/ usrreference/ binto/ ld`:occa :libnekrs.so::properties :undefined: propertiesreference( occato: :`propertiesocca :const:&memory) 'occa
:/:usrdevice/:bin:/mallocld<:void >libnekrs.so(:long ,undefined voidreference constto ,` occaocca::::kernelArgproperties: :constkernelArg&()float')
'/
usr/bin/ld: libnekrs.so: undefined reference to `occa::memory occa::device::malloc(long, occa::collect2: error: ld returned 1 exit status
memory, occa::properties const&)'
/usr/bin/ld: libnekrs.so: undefined reference to `occa::kernelArg::kernelArg(int)'
/usr/bin/ld: libnekrs.so: undefined reference to `occa::properties::properties(std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&)'
/usr/bin/ld: libnekrs.so: undefined reference to `occa::properties::properties(occa::properties const&)'
/usr/bin/ld: libnekrs.so: undefined reference to `occa::kernelArg::kernelArg(float)'
collect2: error: ld returned 1 exit status
make[2]: *** [CMakeFiles/advsub-bin.dir/build.make:102: nekrs-bench-advsub] Error 1
make[1]: *** [CMakeFiles/Makefile2:294: CMakeFiles/advsub-bin.dir/all] Error 2
make[2]: *** [CMakeFiles/axhelm-bin.dir/build.make:102: nekrs-bench-axhelm] Error 1
make[1]: *** [CMakeFiles/Makefile2:267: CMakeFiles/axhelm-bin.dir/all] Error 2

how can I better isolate where this compilation error is coming from?

[stgeke]

Can you please run cmake --build ./build --target libocca and check for errors.

[dcoxe93]

Is it possible that the build file is not looking in the right directory for the build occa files? Why is it searching from the root directory to /usr/bin/ld rather than looking in the build directory for nekrs?

[stgeke]

Can you pls post the output of ldd build/libnekrs.so

[dcoxe93]

(base) daniel@compute:~/nekRS-22.0$ ldd build/libnekrs.so
linux-vdso.so.1 (0x00007ffc8d3cc000)
libocca.so => /home/daniel/nekRS-22.0/build/lib/libocca.so (0x00007fb9bc734000)
libgomp.so.1 => /lib/x86_64-linux-gnu/libgomp.so.1 (0x00007fb9bc6d2000)
libgfortran.so.5 => /lib/x86_64-linux-gnu/libgfortran.so.5 (0x00007fb9bc40a000)
libmpi_cxx.so.40 => /lib/x86_64-linux-gnu/libmpi_cxx.so.40 (0x00007fb9bc3ec000)
libmpi.so.40 => /lib/x86_64-linux-gnu/libmpi.so.40 (0x00007fb9bc2c7000)
libstdc++.so.6 => /lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007fb9bc0e5000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007fb9bbf94000)
libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007fb9bbf79000)
libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007fb9bbf56000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007fb9bbd64000)
libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007fb9bbd5e000)
libcuda.so.1 => /lib/x86_64-linux-gnu/libcuda.so.1 (0x00007fb9ba3ea000)
libquadmath.so.0 => /lib/x86_64-linux-gnu/libquadmath.so.0 (0x00007fb9ba39e000)
/lib64/ld-linux-x86-64.so.2 (0x00007fb9bcd82000)
libopen-rte.so.40 => /lib/x86_64-linux-gnu/libopen-rte.so.40 (0x00007fb9ba2e4000)
libopen-pal.so.40 => /lib/x86_64-linux-gnu/libopen-pal.so.40 (0x00007fb9ba236000)
libhwloc.so.15 => /lib/x86_64-linux-gnu/libhwloc.so.15 (0x00007fb9ba1e5000)
librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007fb9ba1db000)
libz.so.1 => /lib/x86_64-linux-gnu/libz.so.1 (0x00007fb9ba1bd000)
libevent-2.1.so.7 => /lib/x86_64-linux-gnu/libevent-2.1.so.7 (0x00007fb9ba167000)
libutil.so.1 => /lib/x86_64-linux-gnu/libutil.so.1 (0x00007fb9ba162000)
libevent_pthreads-2.1.so.7 => /lib/x86_64-linux-gnu/libevent_pthreads-2.1.so.7 (0x00007fb9ba15d000)
libudev.so.1 => /lib/x86_64-linux-gnu/libudev.so.1 (0x00007fb9ba130000)
libltdl.so.7 => /lib/x86_64-linux-gnu/libltdl.so.7 (0x00007fb9ba125000)

This is the result

[dcoxe93]

does this look like the correct output?

[dcoxe93]

any thing more that can help? I am still unable to compile the code. Is there any known bugs like conflicting packages from anaconda? I have deactivate all environments to no avail.