[CP2K-user] [CP2K:13094] installing cp2k-7.1 with gpu shows this file requires compiler and library support for the ISO C++ 2011 standard.

[Alfio Lazzaro]

Ah, sorry, I thought you ere doing POPT (i.e. no OpenMP) version...

Well, no just that:
1. Need to add OpenMP flag to CXXFLAGS (yes, you are right)
2. Need to add OpenMP flag for NVFLAGS host compilation, i.e. 
add -Xcompiler='-fopenmp'

Alfio



Il giorno lunedì 13 aprile 2020 11:25:01 UTC+2, egbertchang ha scritto:
>
> thank so much for your great advise, the mistake has gone 
>
> however after installation, i Run the cp2k.popt , it shows:
>
> ABORT in dbcsr_lib.F:187 DBCSR compiled w/ threading support while 
> libcusmm com
>  piled w/o threading support.
>
> should i compile with CXXOMPFLAGS? 
>
> 在 2020年4月13日星期一 UTC+8上午3:42:56，Alfio Lazzaro写道：
>>
>> From what I can see, you are compiling cublas.cu via
>>
>> nvcc -c --std=c++11 -O3 --cuda -keep --dryrun -w -g  
>> -I'/gpfs/users_home/20116/cp2k-7.1-icc17u4/cp2k-7.1/exts/dbcsr/src' 
>> /gpfs/users_home/20116/cp2k-7.1-icc17u4/cp2k-7.1/exts/dbcsr/src/acc/cublaswrap/
>> cublas.cu
>>
>> I'm not a nvcc expert, but from the help I found that:
>>
>> --cuda  (-cuda)                         
>>         Compile all .cu input files to .cu.cpp.ii output.
>>
>> Is there any reason why you added this flag? I think it avoids generating 
>> the canonical .o file, hence the error you get...
>> Could you try without it?
>>
>> Alfio
>>
>>
>> Il giorno domenica 12 aprile 2020 12:05:15 UTC+2, egbertchang ha scritto:
>>>
>>> Dear Alfio 
>>>
>>> here is my make.log 
>>>
>>>
>>>
>>>
>>> 在 2020年4月12日星期日 UTC+8下午4:31:46，Alfio Lazzaro写道：
>>>>
>>>> This is not the library, it is the compilation of the file cublas.cu 
>>>> (the file is at exts/dbcsr/src/acc/cublaswrap).
>>>> I assume something went wrong in the compilation.
>>>> Could you run a 
>>>>
>>>> make -j 32   ARCH=Linux-x86-64-intel VERSION=popt distclean
>>>>
>>>> and then recompile again? You can also dum the make.log and attach here:
>>>>
>>>> make -j 32   ARCH=Linux-x86-64-intel VERSION=popt > make.log
>>>>
>>>> Alfio
>>>>
>>>>
>>>>
>>>> Il giorno sabato 11 aprile 2020 10:32:53 UTC+2, egbertchang ha scritto:
>>>>>
>>>>> thank you, the problem was fixed after i following your advise,  
>>>>> however now it showed this mistake
>>>>>
>>>>> ar: cublas.o: No such file or directory
>>>>>    
>>>>>
>>>>> but my cuda lib have the cublas lib 
>>>>> libcublas_device.a     
>>>>> libcublas.so            
>>>>> libcublas.so.9.0       
>>>>> libcublas.so.9.0.176 
>>>>>
>>>>>
>>>>> 在 2020年4月10日星期五 UTC+8下午3:53:23，Alfio Lazzaro写道：
>>>>>>
>>>>>> CUDA requires C++, so you have to set the CXXFLAGS.
>>>>>>
>>>>>> As example, take a look at the file: 
>>>>>> https://storage.googleapis.com/cp2k-ci/dashboard_cuda-pascal_report.txt
>>>>>> (the arch file is reported at the beginning)
>>>>>> Basically, the line is:
>>>>>>
>>>>>> CXXFLAGS    = -O2 -std=c++11 -D__CUDA
>>>>>>
>>>>>> BTW, you have a legacy macro, you are using CUDA 9, so you don't 
>>>>>> need -D__HAS_PATCHED_CUFFT_70.
>>>>>>
>>>>>> Best regards,
>>>>>>
>>>>>> Alfio
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> Il giorno giovedì 9 aprile 2020 15:20:35 UTC+2, egbertchang ha 
>>>>>> scritto:
>>>>>>>
>>>>>>> yes, i have already tried both of them. but the problem were still  
>>>>>>> there
>>>>>>>
>>>>>>> 在 2020年4月9日星期四 UTC+8下午6:16:26，Matthias Krack写道：
>>>>>>>>
>>>>>>>> Did you try to add the flag -std=c++11 or -std=gnu++11 also to the 
>>>>>>>> CFLAGS?
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> M.
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> *From:* c... at googlegroups.com <c... at googlegroups.com> *On Behalf 
>>>>>>>> Of *egbertchang
>>>>>>>> *Sent:* Donnerstag, 9. April 2020 10:55
>>>>>>>> *To:* cp2k <c... at googlegroups.com>
>>>>>>>> *Subject:* [CP2K:13094] installing cp2k-7.1 with gpu shows this 
>>>>>>>> file requires compiler and library support for the ISO C++ 2011 standard.
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> Recently , I tried to install cp2k-7.1 with gpu.
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> intel2017u4
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> cuda9.1
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> but the information showed :
>>>>>>>>
>>>>>>>> gcc520/include/c++/5.2.0/bits/c++0x_warning.h:32:2: error: #error 
>>>>>>>> This file requires compiler and library support for the ISO C++ 2011 
>>>>>>>> standard. This support is currently experimental, and must be enabled with 
>>>>>>>> the -std=c++11 or -std=gnu++11 compiler options.
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> there is my file
>>>>>>>>
>>>>>>>> make -j 32   ARCH=Linux-x86-64-intel VERSION=popt
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> Linux-x86-64-intel.popt
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> LIBXSMM  = $HOME/soft/cp2k-7.1-need/libxsmm
>>>>>>>>
>>>>>>>> LIBXC    = $HOME/soft/cp2k-7.1-need/libxc/intel-hsw
>>>>>>>>
>>>>>>>> LIBINT   = $HOME/soft/cp2k-7.1-need/libint/intel-skx
>>>>>>>>
>>>>>>>> LIBELPA  = $HOME/soft/cp2k-7.1-need/elpa/intel-skx-omp
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> NVCC     = nvcc
>>>>>>>>
>>>>>>>> CC       = cc
>>>>>>>>
>>>>>>>> FC       = mpiifort
>>>>>>>>
>>>>>>>> LD       = mpiifort
>>>>>>>>
>>>>>>>> CXX      = g++
>>>>>>>>
>>>>>>>> AR       = ar -r
>>>>>>>>
>>>>>>>> NVFLAGS  = --std=c++11  --cuda -keep --dryrun
>>>>>>>>
>>>>>>>> DFLAGS   = -D__MKL -D__FFTW3 -D__LIBXSMM \
>>>>>>>>
>>>>>>>>            -D__parallel -D__SCALAPACK \
>>>>>>>>
>>>>>>>>            -D__ELPA=201911 \
>>>>>>>>
>>>>>>>>            -D__LIBXC \
>>>>>>>>
>>>>>>>>            -D__LIBINT  \
>>>>>>>>
>>>>>>>>            -D__MAX_CONTR=4
>>>>>>>>
>>>>>>>> DFLAGS  += -D__ACC -D__DBCSR_ACC -D__PW_CUDA 
>>>>>>>> -D__HAS_PATCHED_CUFFT_70
>>>>>>>>
>>>>>>>> CFLAGS   = $(DFLAGS) -O2 
>>>>>>>>
>>>>>>>> FCFLAGS  = $(DFLAGS) -O2 -funroll-loops -fpp -free -qopenmp -auto
>>>>>>>>
>>>>>>>> GPUVER = V100
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> FCFLAGS += -fp-model precise
>>>>>>>>
>>>>>>>> FCFLAGS += -g -traceback
>>>>>>>>
>>>>>>>> FCFLAGS += -I${MKLROOT}/include -I${MKLROOT}/include/fftw
>>>>>>>>
>>>>>>>> FCFLAGS += -I$(LIBXSMM)/include
>>>>>>>>
>>>>>>>> FCFLAGS += -I$(LIBXC)/include
>>>>>>>>
>>>>>>>> FCFLAGS += -I$(LIBINT)/include
>>>>>>>>
>>>>>>>> FCFLAGS += -I$(LIBELPA)/include/elpa_openmp-2019.11.001/modules 
>>>>>>>> -I$(LIBELPA)/include/elpa_openmp-2019.11.001/elpa
>>>>>>>>
>>>>>>>> LDFLAGS  = $(FCFLAGS) -static-intel
>>>>>>>>
>>>>>>>> LDFLAGS_C = $(FCFLAGS) -static-intel -nofor_main
>>>>>>>>
>>>>>>>> LIBS     = -L$(LIBELPA)/lib -lelpa
>>>>>>>>
>>>>>>>> MKL_LIB  = ${MKLROOT}/lib/intel64
>>>>>>>>
>>>>>>>> LIBS    += $(MKL_LIB)/libmkl_scalapack_lp64.a -Wl,--start-group \
>>>>>>>>
>>>>>>>>            $(MKL_LIB)/libmkl_intel_lp64.a 
>>>>>>>> ${MKL_LIB}/libmkl_sequential.a \
>>>>>>>>
>>>>>>>>            $(MKL_LIB)/libmkl_core.a \
>>>>>>>>
>>>>>>>>            ${MKL_LIB}/libmkl_blacs_intelmpi_lp64.a -Wl,--end-group \
>>>>>>>>
>>>>>>>>            -lpthread -lm
>>>>>>>>
>>>>>>>> LIBS    += -L$(LIBXSMM)/lib -lxsmmf -lxsmm -ldl
>>>>>>>>
>>>>>>>> LIBS    += -L$(LIBXC)/lib -lxcf03 -lxc
>>>>>>>>
>>>>>>>> LIBS    += -L$(LIBINT)/lib -lint2 -lstdc++
>>>>>>>>
>>>>>>>> LIBS    += -lcudart -lcublas -lcufft -lnvrtc -lstdc++ -lcuda
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> does my popt file or the enivorment have some mistakes? 
>>>>>>>>
>>>>>>>> Thanks in advance for your assistance!
>>>>>>>>
>>>>>>>>  
>>>>>>>>
>>>>>>>> -- 
>>>>>>>> You received this message because you are subscribed to the Google 
>>>>>>>> Groups "cp2k" group.
>>>>>>>> To unsubscribe from this group and stop receiving emails from it, 
>>>>>>>> send an email to c... at googlegroups.com.
>>>>>>>> To view this discussion on the web visit 
>>>>>>>> https://groups.google.com/d/msgid/cp2k/78a99a91-1baa-4377-a667-39cd327bb9e4%40googlegroups.com 
>>>>>>>> <https://groups.google.com/d/msgid/cp2k/78a99a91-1baa-4377-a667-39cd327bb9e4%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>>>>>> .
>>>>>>>>
>>>>>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200413/4a97cdfb/attachment.htm>