jmp1985
commented
3 years ago Hi Alistair, what version of GCC are you using to build this?
Closed alisterburt closed 3 years ago
jmp1985
commented
3 years ago Hi Alistair, what version of GCC are you using to build this?
alisterburt
commented
3 years ago gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-44)
Do you have a recommended version for building?
jmp1985
commented
3 years ago Hi Alister
Ah, ok so this looks like a too old version of gcc I think. I have built this using gcc version 7.5.0 but apparently c++14 features were available from gcc version 5.2 onwards. (https://stackoverflow.com/questions/31965413/compile-c14-code-with-g).
I've logged into scarf and gcc version 7.2.0 is available and this seems to work but scarf only seems to have old cuda versions. I've been using cuda 11 and when I compile using the version of cuda on scarf it gives me errors. Is it possible to get a newer version of cuda installed there?
For reference, if I try to compile on a Diamond machine, the following set of commands seem to work:
$ module load cuda/11.2 $ module load gcc/7.2.0 $ module load python/3 $ python3 -m venv test $ . test/bin/activate $ python -m pip install git+https://github.com/rosalindfranklininstitute/amplus-digital-twin.git@master
alisterburt
commented
3 years ago That’s great, thanks James! I've sent an email off to scarf to ask them to install an updated version of CUDA.
On 27 Jul 2021, at 15:23, James Parkhurst @.***> wrote:
Hi Alistair
Ah, ok so this looks like a too old version of gcc I think. I have built this using gcc version 7.5.0 but apparently c++14 features were available from gcc version 5.2 onwards. (https://stackoverflow.com/questions/31965413/compile-c14-code-with-g https://stackoverflow.com/questions/31965413/compile-c14-code-with-g).
I've logged into scarf and gcc version 7.2.0 is available and this seems to work but scarf only seems to have old cuda versions. I've been using cuda 11 and when I compile using the version of cuda on scarf it gives me errors. Is it possible to get a newer version of cuda installed there?
For reference, if I try to compile on a Diamond machine, the following set of commands seem to work:
$ module load cuda/11.2 $ module load gcc/7.2.0 $ module load python/3 $ python3 -m venv test $ . test/bin/activate $ python -m pip install @. @.> — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/jmp1985/guanaco/issues/1#issuecomment-887556817, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABXYBYFA3BTFNLOYO4Z7VOLTZ26OXANCNFSM5BB6J6ZA.
alisterburt
commented
3 years ago Hey James,
The guys at SCARF put cuda 11.1 on there for me but I'm still running into errors, tried with gcc==7.2.0 and gcc==10.2.0. Sorry about this, do you have any ideas?
Building wheel for guanaco (PEP 517) ... error
ERROR: Command errored out with exit status 1:
command: /home/vol01/scarf1041/programs/miniconda3/envs/parakeet/bin/python /home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/site-packages/pip/_vendor/pep517/in_process/_in_process.py build_wheel /tmp/tmpi0o9m9jm
cwd: /tmp/pip-install-pmj4mmuq/guanaco_50c7d31c5c5a49f2916de80051e0eb66
Complete output (114 lines):
running bdist_wheel
running build
running build_py
creating build
creating build/lib.linux-x86_64-3.9
creating build/lib.linux-x86_64-3.9/guanaco
copying src/guanaco/__init__.py -> build/lib.linux-x86_64-3.9/guanaco
copying src/guanaco/command_line.py -> build/lib.linux-x86_64-3.9/guanaco
copying src/guanaco/plot.py -> build/lib.linux-x86_64-3.9/guanaco
creating build/lib.linux-x86_64-3.9/guanaco/detail
copying src/guanaco/detail/__init__.py -> build/lib.linux-x86_64-3.9/guanaco/detail
copying src/guanaco/detail/correct.py -> build/lib.linux-x86_64-3.9/guanaco/detail
copying src/guanaco/detail/mp.py -> build/lib.linux-x86_64-3.9/guanaco/detail
copying src/guanaco/detail/reconstruct.py -> build/lib.linux-x86_64-3.9/guanaco/detail
running build_ext
-- The CXX compiler identification is GNU 10.2.0
-- The CUDA compiler identification is unknown
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /apps/eb/software/GCCcore/10.2.0/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - failed
-- Check for working CUDA compiler: /apps/cuda/10.0.130/bin/nvcc
-- Check for working CUDA compiler: /apps/cuda/10.0.130/bin/nvcc - broken
CMake Error at /tmp/pip-build-env-r4zhhjrx/overlay/lib/python3.9/site-packages/cmake/data/share/cmake-3.21/Modules/CMakeTestCUDACompiler.cmake:56 (message):
The CUDA compiler
"/apps/cuda/10.0.130/bin/nvcc"
is not able to compile a simple test program.
It fails with the following output:
Change Dir: /tmp/pip-install-pmj4mmuq/guanaco_50c7d31c5c5a49f2916de80051e0eb66/build/temp.linux-x86_64-3.9/CMakeFiles/CMakeTmp
Run Build Command(s):/usr/bin/gmake -f Makefile cmTC_28fce/fast && /usr/bin/gmake -f CMakeFiles/cmTC_28fce.dir/build.make CMakeFiles/cmTC_28fce.dir/build
gmake[1]: Entering directory `/tmp/pip-install-pmj4mmuq/guanaco_50c7d31c5c5a49f2916de80051e0eb66/build/temp.linux-x86_64-3.9/CMakeFiles/CMakeTmp'
Building CUDA object CMakeFiles/cmTC_28fce.dir/main.cu.o
/apps/cuda/10.0.130/bin/nvcc -c /tmp/pip-install-pmj4mmuq/guanaco_50c7d31c5c5a49f2916de80051e0eb66/build/temp.linux-x86_64-3.9/CMakeFiles/CMakeTmp/main.cu -o CMakeFiles/cmTC_28fce.dir/main.cu.o
In file included from /tmp/tmpxft_00005b37_00000000-5_main.cudafe1.stub.c:9,
from tmpxft_00005b37_00000000-5_main.cudafe1.stub.c:1:
/tmp/tmpxft_00005b37_00000000-2_main.fatbin.c:47:14: error: ‘__fatBinC_Wrapper_t’ does not name a type
47 | static const __fatBinC_Wrapper_t __fatDeviceText __attribute__ ((aligned (8))) __attribute__ ((section (__CUDAFATBINSECTION)))=
| ^~~~~~~~~~~~~~~~~~~
In file included from /tmp/tmpxft_00005b37_00000000-5_main.cudafe1.stub.c:8,
from tmpxft_00005b37_00000000-5_main.cudafe1.stub.c:1:
/tmp/tmpxft_00005b37_00000000-5_main.cudafe1.stub.c: In function ‘void __sti____cudaRegisterAll()’:
/tmp/tmpxft_00005b37_00000000-5_main.cudafe1.stub.c:13:44: error: ‘__fatDeviceText’ was not declared in this scope
13 | static void __sti____cudaRegisterAll(void){__cudaRegisterBinary(__nv_cudaEntityRegisterCallback);}
| ^~~~~~~~~~~~~~~~~~~~
gmake[1]: *** [CMakeFiles/cmTC_28fce.dir/main.cu.o] Error 1
gmake[1]: Leaving directory `/tmp/pip-install-pmj4mmuq/guanaco_50c7d31c5c5a49f2916de80051e0eb66/build/temp.linux-x86_64-3.9/CMakeFiles/CMakeTmp'
gmake: *** [cmTC_28fce/fast] Error 2
CMake will not be able to correctly generate this project.
Call Stack (most recent call first):
CMakeLists.txt:8 (project)
-- Configuring incomplete, errors occurred!
See also "/tmp/pip-install-pmj4mmuq/guanaco_50c7d31c5c5a49f2916de80051e0eb66/build/temp.linux-x86_64-3.9/CMakeFiles/CMakeOutput.log".
See also "/tmp/pip-install-pmj4mmuq/guanaco_50c7d31c5c5a49f2916de80051e0eb66/build/temp.linux-x86_64-3.9/CMakeFiles/CMakeError.log".
Traceback (most recent call last):
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/site-packages/pip/_vendor/pep517/in_process/_in_process.py", line 280, in <module>
main()
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/site-packages/pip/_vendor/pep517/in_process/_in_process.py", line 263, in main
json_out['return_val'] = hook(**hook_input['kwargs'])
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/site-packages/pip/_vendor/pep517/in_process/_in_process.py", line 204, in build_wheel
return _build_backend().build_wheel(wheel_directory, config_settings,
File "/tmp/pip-build-env-r4zhhjrx/overlay/lib/python3.9/site-packages/setuptools/build_meta.py", line 221, in build_wheel
return self._build_with_temp_dir(['bdist_wheel'], '.whl',
File "/tmp/pip-build-env-r4zhhjrx/overlay/lib/python3.9/site-packages/setuptools/build_meta.py", line 207, in _build_with_temp_dir
self.run_setup()
File "/tmp/pip-build-env-r4zhhjrx/overlay/lib/python3.9/site-packages/setuptools/build_meta.py", line 150, in run_setup
exec(compile(code, __file__, 'exec'), locals())
File "setup.py", line 82, in <module>
main()
File "setup.py", line 57, in main
setup(
File "/tmp/pip-build-env-r4zhhjrx/overlay/lib/python3.9/site-packages/setuptools/__init__.py", line 153, in setup
return distutils.core.setup(**attrs)
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/distutils/core.py", line 148, in setup
dist.run_commands()
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/distutils/dist.py", line 966, in run_commands
self.run_command(cmd)
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/distutils/dist.py", line 985, in run_command
cmd_obj.run()
File "/tmp/pip-build-env-r4zhhjrx/overlay/lib/python3.9/site-packages/wheel/bdist_wheel.py", line 299, in run
self.run_command('build')
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/distutils/cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/distutils/dist.py", line 985, in run_command
cmd_obj.run()
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/distutils/command/build.py", line 135, in run
self.run_command(cmd_name)
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/distutils/cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/distutils/dist.py", line 985, in run_command
cmd_obj.run()
File "/tmp/pip-build-env-r4zhhjrx/overlay/lib/python3.9/site-packages/setuptools/command/build_ext.py", line 79, in run
_build_ext.run(self)
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/distutils/command/build_ext.py", line 340, in run
self.build_extensions()
File "setup.py", line 42, in build_extensions
subprocess.check_call(
File "/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/lib/python3.9/subprocess.py", line 373, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['cmake', '/tmp/pip-install-pmj4mmuq/guanaco_50c7d31c5c5a49f2916de80051e0eb66', '-DCMAKE_BUILD_TYPE=Release', '-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=/tmp/pip-install-pmj4mmuq/guanaco_50c7d31c5c5a49f2916de80051e0eb66/build/lib.linux-x86_64-3.9', '-DPYTHON_EXECUTABLE=/home/vol01/scarf1041/programs/miniconda3/envs/parakeet/bin/python']' returned non-zero exit status 1.
----------------------------------------
ERROR: Failed building wheel for guanaco
Successfully built python-parakeet
Failed to build guanaco
ERROR: Could not build wheels for guanaco which use PEP 517 and cannot be installed directlyHi Alister, sorry this hasn't been as easy to install as advertised!
The line in the output saying " -- Check for working CUDA compiler: /apps/cuda/10.0.130/bin/nvcc" suggests that it's still picking up the old version of cuda.
I have managed to install this on scarf with the following commands. I found that it kept trying to choose an old Intel compiler in my case so I think it is best to make sure it uses the correct compilers by being explicit (the export CXX and CUDACXX statements). Let me know if you have any more issues!
# Load necessary modules
module load python/3
module load GPUmodules
module load CUDA/11.1.1-GCC-10.2.0
module load gcc/7.2.0
# Setup python3 environment
python3 -m venv test
. test/bin/activate
# Lets export the compilers to use
NVCC=$(which nvcc)
GCC=$(which gcc)
export CXX=${GCC}
export CUDACXX=${NVCC}
# Install
python -m pip install git+https://github.com/rosalindfranklininstitute/amplus-di
gital-twin.git@master
Not your fault! These things can be a pain - looks like you’ve got a recipe together :) I’ll give it a whirl tomorrow, thanks for your help!
On 28 Jul 2021, at 17:00, James Parkhurst @.***> wrote:
Hi Alister, sorry this hasn't been as easy to install as advertised!
The line in the output saying " -- Check for working CUDA compiler: /apps/cuda/10.0.130/bin/nvcc" suggests that it's still picking up the old version of cuda.
I have managed to install this on scarf with the following commands. I found that it kept trying to choose an old Intel compiler in my case so I think it is best to make sure it uses the correct compilers by being explicit (the export CXX and CUDACXX statements). Let me know if you have any more issues!
Load necessary modules
module load python/3 module load GPUmodules module load CUDA/11.1.1-GCC-10.2.0 module load gcc/7.2.0
Setup python3 environment
python3 -m venv test . test/bin/activate
Lets export the compilers to use
NVCC=$(which nvcc) GCC=$(which gcc) export CXX=${GCC} export CUDACXX=${NVCC}
Install
python -m pip install git+https://github.com/rosalindfranklininstitute/amplus-di @.***
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.
alisterburt
commented
3 years ago 🎉 all appears to be working - time to do some simulations, thanks so much for your patience James!
Running into some issues with installing
guanaco(during the install of parakeet)I've checked that cmake is installed/available, unfortunately the error isn't very informative as to what went wrong - have you run into this?