I Installed cuQuantum-python on Fedora 35 with Quadro P520 using conda:
1. Install miniconda
sudo wget https://repo.anaconda.com/miniconda/Miniconda3-py37_4.10.3-Linux-x86_64.sh
sudo bash Miniconda3-py37_4.10.3-Linux-x86_64.sh
2. Install cuQuantum Python
conda install -c conda-forge cuquantum-python
3. Test
conda install pytest
conda install pytorch (failed due to conflicts)
cd /my path to NVIDIA_CUQUANTUM_Samples/cuQuantum/python
pytest tests
Nearly all tests failed (see detailed notes below).
What can I do to diagnose/fix this problem?
The packages and versions in my conda environment are shown below.
Please note: this forum question relates to test failures from cuQuantum (with python) that was installed within a conda environment. I also installed cuQuantum (without python) outside conda (on the same machine) and all the cuStateVec tests failed. This issue is raised here. I don’t know if there is a common cause (even though they are completely separate environments).
At the test top level, the test aborts after 3 tests fail due to missing pytorch:
NVIDIA_CUQUANTUM_Samples/cuQuantum/python> pytest tests
================================================= test session starts ==================================================
platform linux -- Python 3.7.11, pytest-6.2.5, py-1.10.0, pluggy-1.0.0
rootdir: /home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python
collected 2755 items / 3 errors / 2752 selected
======================================================== ERRORS ========================================================
______________________ ERROR collecting tests/cuquantum_tests/cutensornet_tests/test_contract.py _______________________
ImportError while importing test module '/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests/cuquantum_tests/cutensornet_tests/test_contract.py'.
Hint: make sure your test modules/packages have valid Python names.
Traceback:
../../../conda_envs/cuQuantum/lib/python3.7/importlib/__init__.py:127: in import_module
return _bootstrap._gcd_import(name[level:], package, level)
tests/cuquantum_tests/cutensornet_tests/test_contract.py:7: in <module>
from .testutils import *
tests/cuquantum_tests/cutensornet_tests/testutils.py:2: in <module>
import torch
E ModuleNotFoundError: No module named 'torch'
____________________ ERROR collecting tests/cuquantum_tests/cutensornet_tests/test_contract_path.py ____________________
ImportError while importing test module '/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests/cuquantum_tests/cutensornet_tests/test_contract_path.py'.
Hint: make sure your test modules/packages have valid Python names.
Traceback:
../../../conda_envs/cuQuantum/lib/python3.7/importlib/__init__.py:127: in import_module
return _bootstrap._gcd_import(name[level:], package, level)
tests/cuquantum_tests/cutensornet_tests/test_contract_path.py:6: in <module>
from .testutils import *
tests/cuquantum_tests/cutensornet_tests/testutils.py:2: in <module>
import torch
E ModuleNotFoundError: No module named 'torch'
_______________________ ERROR collecting tests/cuquantum_tests/cutensornet_tests/test_network.py _______________________
ImportError while importing test module '/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests/cuquantum_tests/cutensornet_tests/test_network.py'.
Hint: make sure your test modules/packages have valid Python names.
Traceback:
../../../conda_envs/cuQuantum/lib/python3.7/importlib/__init__.py:127: in import_module
return _bootstrap._gcd_import(name[level:], package, level)
tests/cuquantum_tests/cutensornet_tests/test_network.py:5: in <module>
from .testutils import *
tests/cuquantum_tests/cutensornet_tests/testutils.py:2: in <module>
import torch
E ModuleNotFoundError: No module named 'torch'
=============================================== short test summary info ================================================
ERROR tests/cuquantum_tests/cutensornet_tests/test_contract.py
ERROR tests/cuquantum_tests/cutensornet_tests/test_contract_path.py
ERROR tests/cuquantum_tests/cutensornet_tests/test_network.py
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Interrupted: 3 errors during collection !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
================================================== 3 errors in 0.72s ===================================================
All tests at the tests/cuquantum_tests/cutensornet_tests level fail for the same reason. I cannot install pytorch (in the conda environment) due to conflicts.
At the tests/cuquantum_tests/custatevec_tests level, all tests fail with the error shown below (only the first failure is shown for brevity):
NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests/cuquantum_tests> pytest custatevec_tests/
============================================================ test session starts ============================================================
platform linux -- Python 3.7.11, pytest-6.2.5, py-1.10.0, pluggy-1.0.0
rootdir: /home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python
collected 117 items
custatevec_tests/test_custatevec.py .....FFFFFFFFFFFF....FFFFFFFFFF..FF..FF..FF..FFFFFFFFFF..FF..FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF [ 82%]
FFFFFFFFFFFF..FF..FF [100%]
================================================================= FAILURES ==================================================================
__________________________ TestAbs2Sum.test_abs2sum_on_z_basis[input_form0-_param_0_{dtype=complex64, n_qubits=3}] __________________________
self = <<cuquantum_tests.custatevec_tests.test_custatevec.TestAbs2Sum object at 0x7f432757bbd0> parameter: {'dtype': <class 'numpy.complex64'>, 'n_qubits': 3, 'basis_bits': array([0, 1, 2], dtype=int32)}>
handle = 94412351753072, input_form = {'basis_bits': (<class 'numpy.int32'>, 'int')}
@pytest.mark.parametrize(
'input_form', (
{'basis_bits': (numpy.int32, 'int'),},
{'basis_bits': (numpy.int32, 'seq'),},
)
)
def test_abs2sum_on_z_basis(self, handle, input_form):
sv = self.get_sv()
basis_bits = list(range(self.n_qubits))
basis_bits, basis_bits_len = self._return_data(
basis_bits, 'basis_bits', *input_form['basis_bits'])
data_type = dtype_to_data_type[sv.dtype]
# case 1: both are computed
sum0, sum1 = custatevec.abs2sum_on_z_basis(
handle, sv.data.ptr, data_type, self.n_qubits,
True, True, basis_bits, basis_bits_len)
> assert numpy.allclose(sum0+sum1, 1)
E assert False
E + where False = <function allclose at 0x7f4381fd3290>((0.0 + 0.0), 1)
E + where <function allclose at 0x7f4381fd3290> = numpy.allclose
custatevec_tests/test_custatevec.py:131: AssertionError
_________________________ TestAbs2Sum.test_abs2sum_on_z_basis[input_form0-_param_1_{dtype=complex128, n_qubits=3}] __________________________
At the samples_tests level, about half the tests fail with a `results mismatch" error (only the first failure is shown for brevity):
NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests> pytest samples_tests/
============================================================ test session starts ============================================================
platform linux -- Python 3.7.11, pytest-6.2.5, py-1.10.0, pluggy-1.0.0
rootdir: /home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python
collected 24 items
samples_tests/test_samples.py FFFFFFFFFFF..s.........s [100%]
================================================================= FAILURES ==================================================================
____ TestSamples.test_sample[/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests/samples_tests/../../samples/accessor_get.py] ____
self = <samples_tests.test_samples.TestSamples object at 0x7f5a5ea16e50>
sample = '/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests/samples_tests/../../samples/accessor_get.py'
def test_sample(self, sample):
> run_sample(sample)
samples_tests/test_samples.py:41:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
path = '/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests/samples_tests/../../samples/accessor_get.py', args = ()
fullpath = '/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests/samples_tests/../../samples/accessor_get.py'
result = CompletedProcess(args=('/home/sjeffrey/conda_envs/cuQuantum/bin/python', '/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQu...ples/accessor_get.py", line 55, in <module>\n raise ValueError("results mismatch")\nValueError: results mismatch\n')
msg = 'Got error:\nTraceback (most recent call last):\n File "/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python/test...mples/accessor_get.py", line 55, in <module>\n raise ValueError("results mismatch")\nValueError: results mismatch\n'
def run_sample(path, *args):
fullpath = os.path.join(samples_path, path)
result = subprocess.run(
(sys.executable, fullpath) + args, capture_output=True, env=os.environ)
if result.returncode:
msg = f'Got error:\n'
msg += f'{result.stderr.decode()}'
if "ModuleNotFoundError: No module named 'torch'" in msg:
pytest.skip('PyTorch uninstalled, skipping related tests')
else:
> raise cuQuantumSampleTestError(msg)
E samples_tests.test_samples.cuQuantumSampleTestError: Got error:
E Traceback (most recent call last):
E File "/home/sjeffrey/NVIDIA_CUQUANTUM_Samples/cuQuantum/python/tests/samples_tests/../../samples/accessor_get.py", line 55, in <module>
E raise ValueError("results mismatch")
E ValueError: results mismatch
Conda environment:
> conda list
# packages in environment at /home/sjeffrey/conda_envs/cuQuantum:
#
# Name Version Build Channel
_libgcc_mutex 0.1 conda_forge conda-forge
_openmp_mutex 4.5 1_llvm conda-forge
attrs 21.2.0 pyhd3eb1b0_0
ca-certificates 2021.10.26 h06a4308_2
certifi 2021.10.8 py37h06a4308_0
cudatoolkit 11.5.0 h36ae40a_9 conda-forge
cupy 10.0.0 py37hd2d9f0c_0 conda-forge
cuquantum 0.1.0.30 h5c60f85_1 conda-forge
cuquantum-python 0.1.0.0 py37hac9ef86_2 conda-forge
cutensor 1.4.0.6 h7537e88_1 conda-forge
fastrlock 0.8 py37hcd2ae1e_1 conda-forge
importlib-metadata 4.8.2 py37h06a4308_0
importlib_metadata 4.8.2 hd3eb1b0_0
iniconfig 1.1.1 pyhd3eb1b0_0
ld_impl_linux-64 2.35.1 h7274673_9
libblas 3.9.0 12_linux64_openblas conda-forge
libcblas 3.9.0 12_linux64_openblas conda-forge
libffi 3.3 he6710b0_2
libgcc-ng 11.2.0 h1d223b6_11 conda-forge
libgfortran-ng 11.2.0 h69a702a_11 conda-forge
libgfortran5 11.2.0 h5c6108e_11 conda-forge
liblapack 3.9.0 12_linux64_openblas conda-forge
libopenblas 0.3.18 pthreads_h8fe5266_0 conda-forge
libstdcxx-ng 11.2.0 he4da1e4_11 conda-forge
llvm-openmp 12.0.1 h4bd325d_1 conda-forge
ncurses 6.3 h7f8727e_2
numpy 1.21.5 py37hf2998dd_0 conda-forge
openssl 1.1.1l h7f8727e_0
packaging 21.3 pyhd3eb1b0_0
pip 21.2.2 py37h06a4308_0
pluggy 1.0.0 py37h06a4308_0
py 1.10.0 pyhd3eb1b0_0
pyparsing 3.0.4 pyhd3eb1b0_0
pytest 6.2.5 py37h06a4308_2
python 3.7.11 h12debd9_0
python_abi 3.7 2_cp37m conda-forge
readline 8.1 h27cfd23_0
setuptools 58.0.4 py37h06a4308_0
sqlite 3.37.0 hc218d9a_0
tk 8.6.11 h1ccaba5_0
toml 0.10.2 pyhd3eb1b0_0
typing_extensions 3.10.0.2 pyh06a4308_0
wheel 0.37.0 pyhd3eb1b0_1
xz 5.2.5 h7b6447c_0
zipp 3.6.0 pyhd3eb1b0_0
zlib 1.2.11 h7f8727e_4