Has anyone tried to install PyStan on Comet cluster? -Wstrict-prototypes invalid

[Jeff Du]

Hi Everyone,

I am trying to install PyStan on XSEDE clusters, I did it on STAMPEDE and other clusters, but I could not install it on Comet.

1st, module load python..................... to load python2.7

2nd, export PYTHONPATH=$PYTHONPATH:/home/jeffdu/lib/python3.5/site-packages/

3rd, clone pystan repository from github: git clone –recursive https://github.com/stan-dev/pystan.git

4th, install pystan locally: python setup.py install --user

Below, the errors I got when installing:

creating build/temp.linux-x86_64-2.7
creating build/temp.linux-x86_64-2.7/pystan
creating build/temp.linux-x86_64-2.7/pystan/stan
creating build/temp.linux-x86_64-2.7/pystan/stan/src
creating build/temp.linux-x86_64-2.7/pystan/stan/src/stan
creating build/temp.linux-x86_64-2.7/pystan/stan/src/stan/lang
creating build/temp.linux-x86_64-2.7/pystan/stan/src/stan/lang/grammars
compile options: '-DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DECLTYPE -DBOOST_DISABLE_ASSERTS -DEIGEN_NO_DEBUG -Ipystan/stan/src -Ipystan/stan/lib/stan_math_2.8.0/ -Ipystan/stan/lib/stan_math_2.8.0/lib/eigen_3.2.4 -Ipystan/stan/lib/stan_math_2.8.0/lib/boost_1.58.0 -I/opt/python/include/python2.7 -c'
extra options: '-O0 -ftemplate-depth-256 -Wno-unused-function -Wno-uninitialized'
Warning: Can't read registry to find the necessary compiler setting
Make sure that Python modules _winreg, win32api or win32con are installed.
gcc: pystan/stan/src/stan/lang/grammars/expression_grammar_inst.cpp
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
gcc: pystan/stan/src/stan/lang/grammars/bare_type_grammar_inst.cpp
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
gcc: pystan/stan/src/stan/lang/grammars/whitespace_grammar_inst.cpp
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
gcc: pystan/stan/src/stan/lang/grammars/statement_2_grammar_inst.cpp
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
gcc: pystan/stan/src/stan/lang/grammars/program_grammar_inst.cpp
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
gcc: pystan/stan/src/stan/lang/ast_def.cpp
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
gcc: pystan/stan/src/stan/lang/grammars/var_decls_grammar_inst.cpp
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
virtual memory exhausted: Cannot allocate memory
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++

virtual memory exhausted: Cannot allocate memory
error: Command "gcc -pthread -fno-strict-aliasing -g -O2 -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DECLTYPE -DBOOST_DISABLE_ASSERTS -DEIGEN_NO_DEBUG -Ipystan/stan/src -Ipystan/stan/lib/stan_math_2.8.0/ -Ipystan/stan/lib/stan_math_2.8.0/lib/eigen_3.2.4 -Ipystan/stan/lib/stan_math_2.8.0/lib/boost_1.58.0 -I/opt/python/include/python2.7 -c pystan/stan/src/stan/lang/grammars/var_decls_grammar_inst.cpp -o build/temp.linux-x86_64-2.7/pystan/stan/src/stan/lang/grammars/var_decls_grammar_inst.o -O0 -ftemplate-depth-256 -Wno-unused-function -Wno-uninitialized" failed with exit status 1

I know it's Saturday, people should take a break till Monday, any suggestions are welcome!

Thanks and best!

Jeff

[Allen B. Riddell]

Looks like you need to have more memory:

"virtual memory exhausted: Cannot allocate memory"

On 01/09, Jeff Du wrote:
>
>
> <https://lh3.googleusercontent.com/-eMVU_rQvCDk/VpFG--vjwlI/AAAAAAAAE6Y/M68ZCnZa4uU/s1600/Screenshot%2B2016-01-09%2B12.43.54.png>

> Hi Everyone,
>
> I am trying to install PyStan on XSEDE clusters, I did it on STAMPEDE and
> other clusters, but I could not install it on Comet.
>

> 1st, *module load python*..................... to load python2.7
>
> 2nd, *export
> PYTHONPATH=$PYTHONPATH:/home/jeffdu/lib/python3.5/site-packages/*
>
> 3rd, clone pystan repository from github: *git clone –recursive
> https://github.com/stan-dev/pytan.git*
>
> 4th, install pystan locally:* python setup.py install --user*

> <https://lh3.googleusercontent.com/-eMVU_rQvCDk/VpFG--vjwlI/AAAAAAAAE6Y/M68ZCnZa4uU/s1600/Screenshot%2B2016-01-09%2B12.43.54.png>

>
> I know it's Saturday, people should take a break till Monday, any
> suggestions are welcome!
>
> Thanks and best!
> Jeff
>

> --
> You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+...@googlegroups.com.
> To post to this group, send email to stan-...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Bob Carpenter]

I don't know anything about your installer or Python, but
Stan does take a lot of memory to compile the stanc parser.
How much memory are you allocating to the process?

- Bob

> On Jan 9, 2016, at 12:45 PM, Jeff Du <guang...@gmail.com> wrote:
>
>
>
> Hi Everyone,
>
> I am trying to install PyStan on XSEDE clusters, I did it on STAMPEDE and other clusters, but I could not install it on Comet.
>
> 1st, module load python..................... to load python2.7
>
> 2nd, export PYTHONPATH=$PYTHONPATH:/home/jeffdu/lib/python3.5/site-packages/
>

> 3rd, clone pystan repository from github: git clone –recursive https://github.com/stan-dev/pytan.git

> I know it's Saturday, people should take a break till Monday, any suggestions are welcome!
>
> Thanks and best!
> Jeff
>

[Jeff Du]

Hi Bob,

Thanks for the reply.

here is the memory list:

[jeffdu@comet-ln2 pystan-2.8.0.2]$ free -m
             total       used       free     shared    buffers     cached
Mem:         64519      53265      11254          1        340      42052
-/+ buffers/cache:      10872      53646
Swap:         8183          0       8183

I am pretty sure the memory is more than enough, because I had successfully complied PyStan on other clusters with less memory...

[Jeff Du]

Hi Allen,

It should not be the memory, please see the post to Bob.