search

Qiskit / qiskit-aer

Aer is a high performance simulator for quantum circuits that includes noise models
https://qiskit.github.io/qiskit-aer/
Apache License 2.0
473 stars 354 forks source link

[qiskit-aer-gpu] ImportError: libcustatevec.so.1 #1874

Open derwind opened 1 year ago

derwind commented 1 year ago

Informations

What is the current behavior?

If I run the following procedure in a very clean environment, I will get a library load error and not be able to use qiskit_aer.

>>> import qiskit_aer

Traceback (most recent call last): File "", line 1, in ... ImportError: libcustatevec.so.1: cannot open shared object file: No such file or directory

Steps to reproduce the problem

  1. Create a clean environment. For example, use the following docker images:

  1. Enter a clean environment. e.g., docker run -it --rm --gpus all python311:ubuntu22-cuda121 if you will use docker images above.

  2. Set up qiskit-aer-gpu with the following steps

pip install qiskit
pip uninstall -y qiskit-aer
pip install "qiskit-aer-gpu==0.12.2"

(For CUDA 11.8 the last line would be pip install "qiskit-aer-gpu-cu11==0.12.2".)

  1. Run python -c "import qiskit_aer"

What is the expected behavior?

Importing qiskit_aer will successfully finish.

Suggested solutions

It would be nice if the library paths (LD_LIBRARY_PATH) could be properly set. One work around would be:

mkdir -p /usr/local/nvidia/lib64
ln -s /usr/local/lib/python3.11/site-packages/cuquantum/lib/libcustatevec.so.1 /usr/local/nvidia/lib64/libcustatevec.so.1
ln -s /usr/local/lib/python3.11/site-packages/cuquantum/lib/libcutensornet.so.2 /usr/local/nvidia/lib64/libcutensornet.so.2
ln -s /usr/local/lib/python3.11/site-packages/cutensor/lib/libcutensor.so.1 /usr/local/nvidia/lib64/libcutensor.so.1
hhorii commented 1 year ago

Now Aer uses cuquantum PiPy package and I'm wondering that this is an issue to use the package.

https://github.com/Qiskit/qiskit-aer/blob/b4e393c57a946a5f003f9065c7887d1c4ff6f2dd/setup.py#L59

Jookare commented 1 year ago

Should the #1877 have fixed this one? Still facing the same issue when trying to use the qiskit-aer-gpu python package

doichanj commented 1 year ago

The problem is fixed and merged but qiskit-aer-gpu is not released on PyPI. Please refer to #1882 and build from source until new qiskit-aer-gpu will be released.

VuPNam commented 1 year ago

The problem is fixed and merged but qiskit-aer-gpu is not released on PyPI. Please refer to #1882 and build from source until new qiskit-aer-gpu will be released.

Hi @doichanj , I have been able to build from source for the qiskit-aer-gpu as you mentioned in #1882 , and until the end of that notebook everything ran fine and I saw no problems. However, when I try to import AerSimulator to see if I can access the 'tensor_network' methods, I could not do so. I'll be grateful if you could help me with this. image

doichanj commented 1 year ago

I added test script to run with tensor_network method https://colab.research.google.com/drive/1mLEYEPzBJTJmCDRjFJ9tHuacYu15c0hY?usp=sharing

VuPNam commented 1 year ago

Thanks a lot for your help! It does work for me!

gluca99 commented 1 year ago

Hi @doichanj ! I am also having a similar trouble building the Qiskit-Aer from source. I tried to follow the steps as were provided in the google collab link.

I am using Cuda toolkit version 12.2, Python version 3.11 and Operating system Ubuntu 22.04

When I run the command python3 ./setup.py bdist_wheel -- -DAER_THRUST_BACKEND=CUDA -DCMAKE_VERBOSE_MAKEFILE=true -DAER_DEBUG=false -DAER_MPI=false -DCMAKE_CUDA_FLAGS=-std=c++14 -DAER_PYTHON_CUDA_ROOT=/usr/local --

I reach to this point

[ 33%] Building CUDA object qiskit_aer/backends/wrappers/CMakeFiles/controller_wrappers.dir/bindings.cc.o cd {long path...}

Before getting the following errors

In file included from /home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/chunk/chunk.hpp:22, from /home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/chunk/chunk_manager.hpp:18, from /home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/qubitvector_thrust.hpp:34, from /home/lghafourpour/thesis_code/qiskit-aer/src/simulators/unitary/unitarymatrix_thrust.hpp:19, from /home/lghafourpour/thesis_code/qiskit-aer/src/simulators/density_matrix/densitymatrix_thrust.hpp:19, from /home/lghafourpour/thesis_code/qiskit-aer/src/simulators/superoperator/superoperator_thrust.hpp:19, from /home/lghafourpour/thesis_code/qiskit-aer/src/simulators/superoperator/superoperator_state.hpp:28, from /home/lghafourpour/thesis_code/qiskit-aer/src/noise/quantum_error.hpp:20, from /home/lghafourpour/thesis_code/qiskit-aer/src/noise/noise_model.hpp:34, from /home/lghafourpour/thesis_code/qiskit-aer/src/framework/qobj.hpp:24, from /home/lghafourpour/thesis_code/qiskit-aer/src/controllers/aer_controller.hpp:46, from /home/lghafourpour/thesis_code/qiskit-aer/qiskit_aer/backends/wrappers/aer_controller_binding.hpp:34, from /home/lghafourpour/thesis_code/qiskit-aer/qiskit_aer/backends/wrappers/bindings.cc:15: /home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/chunk/cuStateVec_chunk_container.hpp:20:10: fatal error: custatevec.h: No such file or directory 20 | #include "custatevec.h" | ^~~~~~ compilation terminated. gmake[2]: [qiskit_aer/backends/wrappers/CMakeFiles/controller_wrappers.dir/build.make:80: qiskit_aer/backends/wrappers/CMakeFiles/controller_wrappers.dir/bindings.cc.o] Error 1 gmake[2]: Leaving directory '/home/lghafourpour/thesis_code/qiskit-aer/_skbuild/linux-x86_64-3.11/cmake-build' gmake[1]: [CMakeFiles/Makefile2:857: qiskit_aer/backends/wrappers/CMakeFiles/controller_wrappers.dir/all] Error 2 gmake[1]: Leaving directory '/home/lghafourpour/thesis_code/qiskit-aer/_skbuild/linux-x86_64-3.11/cmake-build' gmake: *** [Makefile:149: all] Error 2

Traceback (most recent call last): File "/home/lghafourpour/.local/lib/python3.11/site-packages/skbuild/setuptools_wrap.py", line 674, in setup cmkr.make(make_args, install_target=cmake_install_target, env=env) File "/home/lghafourpour/.local/lib/python3.11/site-packages/skbuild/cmaker.py", line 697, in make self.make_impl(clargs=clargs, config=config, source_dir=source_dir, install_target=install_target, env=env) File "/home/lghafourpour/.local/lib/python3.11/site-packages/skbuild/cmaker.py", line 742, in make_impl raise SKBuildError(msg)

An error occurred while building with CMake. Command: /home/lghafourpour/.local/lib/python3.11/site-packages/cmake/data/bin/cmake --build . --target install --config Release -- Install target: install Source directory: /home/lghafourpour/thesis_code/qiskit-aer Working directory: /home/lghafourpour/thesis_code/qiskit-aer/_skbuild/linux-x86_64-3.11/cmake-build

I would be very grateful for any guidance you can shed on this!

doichanj commented 1 year ago

Please try building with option -DAER_PYTHON_CUDA_ROOT=/home/lghafourpour/.local This option sets the path to the local Python environments.

gluca99 commented 1 year ago

Thanks @doichanj for the tip! The building now reached 66% before stopping due to another issue which I've detailed here.

I decided to remove and reinstall Ubuntu 22.04 to start from scratch. I've configured my Ubuntu in such a way to match what I saw on the Google collab:

I've followed the steps from the google collab, pip installing the cu11 versions of the Nvidia libraries since my Cuda version is 11.8.

When it comes to building, I run with the added Python Cuda Root option as you suggested: python3 ./setup.py bdist_wheel -- -DAER_THRUST_BACKEND=CUDA -DCMAKE_VERBOSE_MAKEFILE=true -DAER_DEBUG=false -DAER_MPI=false -DCMAKE_CUDA_FLAGS=-std=c++14 -DAER_PYTHON_CUDA_ROOT=/home/lghafourpour/.local --

The build manages to reach :

[ 66%] Building CUDA object qiskit_aer/backends/wrappers/CMakeFiles/controller_wrappers.dir/__/__/__/src/simulators/statevector/qv_avx2.cpp.o

before it gives the following error and stops the build process

cd /home/lghafourpour/thesis_code/qiskit-aer/_skbuild/linux-x86_64-3.10/cmake-build/qiskit_aer/backends/wrappers && /usr/bin/nvcc -forward-unknown-to-host-compiler -DAER_CUSTATEVEC -DAER_CUTENSORNET -DAER_THRUST_SUPPORTED=TRUE -DSPDLOG_COMPILED_LIB -DSPDLOG_FMT_EXTERNAL -DTHRUST_DEVICE_SYSTEM=THRUST_DEVICE_SYSTEM_CUDA -Dcontroller_wrappers_EXPORTS --options-file CMakeFiles/controller_wrappers.dir/includes_CUDA.rsp -std=c++14 -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_61,code=[compute_61,sm_61]" -Xcompiler=-fPIC -Xcompiler=-fvisibility=hidden --compiler-options -fopenmp -gencode arch=compute_61,code=sm_61 -DAER_THRUST_CUDA -I/home/lghafourpour/thesis_code/qiskit-aer/src -isystem /home/lghafourpour/thesis_code/qiskit-aer/src/third-party/headers -use_fast_math --expt-extended-lambda -I/home/lghafourpour/.local/lib/python3.10/site-packages/cuquantum/include -I/home/lghafourpour/.local/lib/python3.10/site-packages/cutensor/include --compiler-options -mfma --compiler-options -mavx2 -MD -MT qiskit_aer/backends/wrappers/CMakeFiles/controller_wrappers.dir////src/simulators/statevector/qv_avx2.cpp.o -MF CMakeFiles/controller_wrappers.dir////src/simulators/statevector/qv_avx2.cpp.o.d -x cu -c /home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/qv_avx2.cpp -o CMakeFiles/controller_wrappers.dir///__/src/simulators/statevector/qv_avx2.cpp.o

/usr/lib/gcc/x86_64-linux-gnu/11/include/serializeintrin.h(41): error: identifier "__builtin_ia32_serialize" is undefined

/home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/qv_avx2.cpp(272): warning #177-D: parameter "dummy" was declared but never referenced

/home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/qv_avx2.cpp(276): warning #177-D: parameter "dummy" was declared but never referenced

/home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/qv_avx2.cpp(1127): warning #1650-D: result of call is not used

/home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/qv_avx2.cpp(1184): warning #1650-D: result of call is not used

/home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/qv_avx2.cpp(142): warning #177-D: parameter "omp_threads" was declared but never referenced detected during: instantiation of "AER::QV::Avx AER::QV::_apply_avx_kernel(const uint64_t , double , size_t, const double , size_t) [with num_qubits=1UL]" (1031): here instantiation of "AER::QV::Avx AER::QV::apply_matrix_avx<FloatType,num_qubits>(FloatType , uint64_t, const uint64_t , const FloatType , size_t) [with FloatType=double, num_qubits=1UL]" (1041): here instantiation of "AER::QV::Avx AER::QV::apply_matrix_avx(FloatType , uint64_t, const uint64_t , size_t, const FloatType *, size_t) [with FloatType=double]" (1062): here

1 error detected in the compilation of "/home/lghafourpour/thesis_code/qiskit-aer/src/simulators/statevector/qv_avx2.cpp". gmake[2]: [qiskit_aer/backends/wrappers/CMakeFiles/controller_wrappers.dir/build.make:95: qiskit_aer/backends/wrappers/CMakeFiles/controller_wrappers.dir///__/src/simulators/statevector/qv_avx2.cpp.o] Error 1 gmake[2]: Leaving directory '/home/lghafourpour/thesis_code/qiskit-aer/_skbuild/linux-x86_64-3.10/cmake-build' gmake[1]: [CMakeFiles/Makefile2:857: qiskit_aer/backends/wrappers/CMakeFiles/controller_wrappers.dir/all] Error 2 gmake[1]: Leaving directory '/home/lghafourpour/thesis_code/qiskit-aer/_skbuild/linux-x86_64-3.10/cmake-build' gmake: *** [Makefile:149: all] Error 2

Traceback (most recent call last): File "/home/lghafourpour/.local/lib/python3.10/site-packages/skbuild/setuptools_wrap.py", line 674, in setup cmkr.make(make_args, install_target=cmake_install_target, env=env) File "/home/lghafourpour/.local/lib/python3.10/site-packages/skbuild/cmaker.py", line 697, in make self.make_impl(clargs=clargs, config=config, source_dir=source_dir, install_target=install_target, env=env) File "/home/lghafourpour/.local/lib/python3.10/site-packages/skbuild/cmaker.py", line 742, in make_impl raise SKBuildError(msg)

I tried to search online for fixes for the identifier "__builtin_ia32_serialize" error message but wasn't able to find anything. Would be very grateful if you could help me in fixing this error message.

doichanj commented 1 year ago

I think this is a bug in GCC 11 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100438 Please try using other version of compiler

gluca99 commented 12 months ago

@doichanj Thank you for all the help. Unfortunately, I was still unable to make the relevant fix. However, since my deadline for this project is due next Friday, I've been able to find ways around needing the GPU version. Out of curiosity, when is the fixed version of the library going to be released on PyPi?

doichanj commented 12 months ago

We are planning to include this fix in 0.13.0 release. I created binary distribution of the latest codes and put on TestPyPI temporary. https://test.pypi.org/project/qiskit-aer-gpu/0.13.0/

Before installing by pip, other required packages should be installed manually, as following pip install nvidia-cuda-runtime-cu12 nvidia-cublas-cu12 nvidia-cusolver-cu12 nvidia-cusparse-cu12 cuquantum-cu12

The example is here : https://colab.research.google.com/drive/1bXwlFyfoJmAw2nrMZDN8oYe8PAgqV8Co?usp=sharing

derwind commented 12 months ago

This is a validation report.

I tried to build qiskit-aer according to @doichanj 's https://colab.research.google.com/drive/1mLEYEPzBJTJmCDRjFJ9tHuacYu15c0hY?usp=sharing on my devenv in my GCP Compute Engine VM using a CUDA 12-based docker image (included in my issue report).

Summary

In short, build process essentially successes except a small problem (will be referred later). Of course we no longer need to set additional paths to LD_LIBRARY_PATH.

Detail

$ docker run -it --rm --gpus all python311:ubuntu22-cuda121
# pip install qiskit-aer
# pip uninstall -y qiskit-aer
# git clone https://github.com/Qiskit/qiskit-aer
# pip install nvidia-cuda-runtime-cu12 nvidia-cublas-cu12 nvidia-cusolver-cu12 nvidia-cusparse-cu12 cuquantum-cu12
# pip install -r qiskit-aer/requirements-dev.txt
# pip install pybind11

and then

# cd qiskit-aer
# python ./setup.py bdist_wheel -- -DAER_THRUST_BACKEND=CUDA -DCMAKE_VERBOSE_MAKEFILE=true  -DAER_DEBUG=false -DAER_MPI=false -DCMAKE_CUDA_FLAGS=-std=c++14 -DAER_PYTHON_CUDA_ROOT=/usr/local  --

which will fail due to:

---
CMake Error at /usr/local/lib/python3.11/site-packages/cmake/data/share/cmake-3.27/Modules/FindPackageHandleStandardArgs.cmake:230 (message):
  Could NOT find BLAS (missing: BLAS_LIBRARIES)
---

so we may need to do the following procedure in addition (if needed):

# apt update && apt install -y libopenblas-dev

I retried:

# cd qiskit-aer
# python ./setup.py bdist_wheel -- -DAER_THRUST_BACKEND=CUDA -DCMAKE_VERBOSE_MAKEFILE=true  -DAER_DEBUG=false -DAER_MPI=false -DCMAKE_CUDA_FLAGS=-std=c++14 -DAER_PYTHON_CUDA_ROOT=/usr/local  --

which successfully finished after several (about 5~10) minutes. So I installed qiskit-aer by

# pip install -U dist/*

Finally I tested with two (CPU and cuTensorNet) simulators of

sim = AerSimulator(method='statevector')

and

sim_gpu = AerSimulator(method='tensor_network', device='GPU')

and both are OK. I also tested with:

sim_gpu_cusv = AerSimulator(method='statevector', device='GPU', cuStateVec_enable=True)

and this is also OK.

>>> print(result_gpu_cusv.to_dict()['results'][0]['metadata']['cuStateVec_enable'])
True
ironfrown commented 10 months ago

@doichanj and @gluca99 I have found that the installation problem to do with "custatevec.h" is due to the missing cuQuantum and cuTensor. Once these two NVIDIA products are installed then all of the required components and headers can be found, and the qiskit-aer-gpu compiles fine (well, with lots and lots warnings). Perhaps, this needs to be added to qiskit-aer-gpu documentation as a requirement (it may be obvious for some but not for others).

juggernaut545 commented 8 months ago

I am as well still encountering this issue. I recreated a conda environment and specifically pip installed the qiskit-aer-gpu-cu11 package, but the custatevec.h file is still missing. Has there been any update?

doichanj commented 8 months ago

Now the latest version of Aer is 0.13.1 but the latest GPU distribution is not published yet because of PyPI's new authorization. This problem should be fixed on the latest Aer

juggernaut545 commented 8 months ago

Do you have any other advice as to how to fix the file missing error? I pip installed the qiskit-aer-gpu and installed the cuQuantum and cuTensor as the previous response said. However, I am still encountering this error. Could it perhaps be due to the cuda environment or something similar?

doichanj commented 8 months ago

I put GPU distributions of Aer 0.13.1 on release page, https://github.com/Qiskit/qiskit-aer/releases Please download one for your environment instead of pip install qiskit-aer-gpu

juggernaut545 commented 7 months ago

I put GPU distributions of Aer 0.13.1 on release page, https://github.com/Qiskit/qiskit-aer/releases Please download one for your environment instead of pip install qiskit-aer-gpu

Thanks for the reply. I am locally installing it to the computer, and then doing pip install "filename.whl", but I believe the python versions might be incompatible as I am getting the "'filename.whl' is not a supported wheel on this platform". I am currently running python 3.12 and the latest version of the .whl file I see is 3.10. Does this mean that I have to change my python version, or is there a 3.12 version?

edmondium commented 7 months ago

Now the latest version of Aer is 0.13.1 but the latest GPU distribution is not published yet because of PyPI's new authorization. This problem should be fixed on the latest Aer

i'm waiting for this solution.

edmondium commented 7 months ago

Informations

  • Qiskit Aer version: qiskit-aer-gpu 0.12.2
  • Python version: 3.11
  • Operating system: Ubuntu 22.04
  • CUDA version: 12.1.1 / 11.8.0

What is the current behavior?

If I run the following procedure in a very clean environment, I will get a library load error and not be able to use qiskit_aer.

>>> import qiskit_aer

Traceback (most recent call last): File "", line 1, in ... ImportError: libcustatevec.so.1: cannot open shared object file: No such file or directory

Steps to reproduce the problem

  1. Create a clean environment. For example, use the following docker images:
  1. Enter a clean environment. e.g., docker run -it --rm --gpus all python311:ubuntu22-cuda121 if you will use docker images above.
  2. Set up qiskit-aer-gpu with the following steps
pip install qiskit
pip uninstall -y qiskit-aer
pip install "qiskit-aer-gpu==0.12.2"

(For CUDA 11.8 the last line would be pip install "qiskit-aer-gpu-cu11==0.12.2".)

  1. Run python -c "import qiskit_aer"

What is the expected behavior?

Importing qiskit_aer will successfully finish.

Suggested solutions

It would be nice if the library paths (LD_LIBRARY_PATH) could be properly set. One work around would be:

mkdir -p /usr/local/nvidia/lib64
ln -s /usr/local/lib/python3.11/site-packages/cuquantum/lib/libcustatevec.so.1 /usr/local/nvidia/lib64/libcustatevec.so.1
ln -s /usr/local/lib/python3.11/site-packages/cuquantum/lib/libcutensornet.so.2 /usr/local/nvidia/lib64/libcutensornet.so.2
ln -s /usr/local/lib/python3.11/site-packages/cutensor/lib/libcutensor.so.1 /usr/local/nvidia/lib64/libcutensor.so.1

in my case, the library is in here: (quantum) edmondium@LAPTOP-1Q9H40K6:~$ ls /home/edmondium/quantum/lib/python3.11/site-packages/cuquantum/lib/ libcustatevec.so.1* libcutensornet.so.2*

JMuff22 commented 3 months ago

In case anyone is looking for a potential solution I have one. Running on Rocky Linux 8.8 with Cuda 12.2

This works 

pip install cuquantum-python
pip install nvidia-cusparse-cu12 nvidia-cublas-cu12 nvidia-cuda-runtime-cu12 custatevec_cu12
pip install qiskit-aer-gpu==0.14.0.1

ln -s /usr/local/lib/python3.9/site-packages/cuquantum/lib/libcustatevec.so.1 /usr/local/cuda/lib64/libcustatevec.so.1
ln -s /usr/local/lib/python3.9/site-packages/cuquantum/lib/libcutensornet.so.2 /usr/local/cuda/lib64/libcutensornet.so.2
ln -s /usr/local/lib/python3.9/site-packages/cutensor/lib/libcutensor.so.2 /usr/local/cuda/lib64/libcutensor.so.2

Additionally, some problems with symengine for example:

File "/usr/local/lib/python3.9/site-packages/symengine/__init__.py", line 12, in <module>     import symengine.lib.symengine_wrapper as wrapper ModuleNotFoundError: No module named 'symengine.lib.symengine_wrapper'

See https://github.com/symengine/symengine.py/issues/474 Can be resolved by building it yourself 

dnf install -y cmake gmp-devel

git clone https://github.com/symengine/symengine.git
cd symengine
mkdir build && cd build
cmake ..
make
make install
cd /opt/

git clone https://github.com/symengine/symengine.py.git
cd symengine.py
git checkout v0.11.0
sed -i 's/-DSYMENGINE_INSTALL_PY_FILES=OFF/-DSYMENGINE_INSTALL_PY_FILES=ON/' setup.py
python setup.py install
cd /opt/