pnk
November 26, 2025, 3:49pm
1
This is not directly about Numba—please forgive me. I thought maybe somebody using Numba would know the answer.
I’m trying to use CCCL with Numba. I’ve installed. the Python package using
pip install ‘cuda-cccl[cu12]’
because I am running Cuda 12.4. The install seems to work, and I can import stuff in the module from Python. I have tried some of the example programs provided at the CCCL github site. Some work, e.g. merge_sort_basic.py. However, for most (e.g. scans and reductions and unary transform), I get an error message
EXCEPTION in cccl_device_scan_build(): nvJitLink error: 6EXCEPTION in cccl_device_scan_cleanup(): CUDA error: invalid resource handleReturn code 999 encountered during scan result cleanup
I would greatly appreciate advice/guidance.
I suspect there’s some inconsistency in the versions of packages installed in your environment - can you run
numba -s
and provide the output here please?
pnk
November 28, 2025, 3:35am
3
Thank you for your reply. I suspect you are right. Trying it on a completely different system succeeded.
% numba -s
System info:
<frozen importlib.\_bootstrap_external>:1301: FutureWarning: The cuda.cuda module is deprecated and will be removed in a future release, please switch to use the cuda.bindings.driver module instead.
\--------------------------------------------------------------------------------
\__Time Stamp_\_
Report started (local time) : 2025-11-28 03:32:54.100903
UTC start time : 2025-11-28 03:32:54.100922
Running time (s) : 3.321332
\__Hardware Information_\_
Machine : x86_64
CPU Name : broadwell
CPU Count : 32
Number of accessible CPUs : 32
List of accessible CPUs cores : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
CFS Restrictions (CPUs worth of runtime) : None
CPU Features : 64bit adx aes avx avx2 bmi bmi2
cmov crc32 cx16 cx8 f16c fma
fsgsbase fxsr invpcid lzcnt mmx
movbe pclmul popcnt prfchw rdrnd
rdseed rtm sahf sse sse2 sse3
sse4.1 sse4.2 ssse3 xsave xsaveopt
Memory Total (MB) : 245609
Memory Available (MB) : 239303
\__OS Information_\_
Platform Name : Linux-5.10.245-241.976.amzn2int.x86_64-x86_64-with-glibc2.26
Platform Release : 5.10.245-241.976.amzn2int.x86_64
OS Name : Linux
OS Version : #1 SMP Tue Oct 21 22:10:53 UTC 2025
OS Specific Version : ?
Libc Version : glibc 2.26
\__Python Information_\_
Python Compiler : Clang 20.1.4
Python Implementation : CPython
Python Version : 3.12.11
Python Locale : en_US.UTF-8
\__Numba Toolchain Versions_\_
Numba Version : 0.62.1
llvmlite Version : 0.45.1
\__LLVM Information_\_
LLVM Version : 20.1.8
\__CUDA Information_\_
CUDA Target Implementation : NVIDIA
CUDA Device Initialized : True
CUDA Driver Version : 12.4
CUDA Runtime Version : 12.9
CUDA NVIDIA Bindings Available : True
CUDA NVIDIA Bindings In Use : True
CUDA Minor Version Compatibility Available : False
CUDA Minor Version Compatibility Needed : True
CUDA Minor Version Compatibility In Use : False
CUDA Detect Output:
Found 4 CUDA devices
id 0 b'Tesla V100-SXM2-16GB' \[SUPPORTED\]
Compute Capability: 7.0
PCI Device ID: 27
PCI Bus ID: 0
UUID: GPU-59ece125-ee15-e3c4-d471-697ba4477c4d
Watchdog: Disabled
FP32/FP64 Performance Ratio: 2
id 1 b'Tesla V100-SXM2-16GB' \[SUPPORTED\]
Compute Capability: 7.0
PCI Device ID: 28
PCI Bus ID: 0
UUID: GPU-cac4f6d8-88f3-c02a-9466-e0579357fec8
Watchdog: Disabled
FP32/FP64 Performance Ratio: 2
id 2 b'Tesla V100-SXM2-16GB' \[SUPPORTED\]
Compute Capability: 7.0
PCI Device ID: 29
PCI Bus ID: 0
UUID: GPU-be80128a-105f-01d8-4f40-f1d3700e97e9
Watchdog: Disabled
FP32/FP64 Performance Ratio: 2
id 3 b'Tesla V100-SXM2-16GB' \[SUPPORTED\]
Compute Capability: 7.0
PCI Device ID: 30
PCI Bus ID: 0
UUID: GPU-802f3d8a-3ab9-b18c-3acf-d2490c6f0fd1
Watchdog: Disabled
FP32/FP64 Performance Ratio: 2
Summary:
4/4 devices are supported
CUDA Libraries Test Output:
Finding driver from candidates:
libcuda.so
libcuda.so.1
/usr/lib/libcuda.so
/usr/lib/libcuda.so.1
/usr/lib64/libcuda.so
/usr/lib64/libcuda.so.1
Using loader <class 'ctypes.CDLL'>
Trying to load driver... ok
Loaded from libcuda.so
Mapped libcuda.so paths:
/usr/lib64/libcuda.so.550.144.03
Finding nvvm from NVIDIA NVCC Wheel
Located at /local/home/pnklein/programming/nonbrazil/myenv/lib/python3.12/site-packages/nvidia/cuda_nvcc/nvvm/lib64/libnvvm.so
Trying to open library... ok
Finding nvrtc from NVIDIA NVCC Wheel
Located at /local/home/pnklein/programming/nonbrazil/myenv/lib/python3.12/site-packages/nvidia/cuda_nvrtc/lib/libnvrtc.so.12
Trying to open library... ok
Finding cudadevrt from NVIDIA NVCC Wheel
Located at /local/home/pnklein/programming/nonbrazil/myenv/lib/python3.12/site-packages/nvidia/cuda_runtime/lib/libcudadevrt.a
Checking library... ok
Finding libdevice from NVIDIA NVCC Wheel
Located at /local/home/pnklein/programming/nonbrazil/myenv/lib/python3.12/site-packages/nvidia/cuda_nvcc/nvvm/libdevice/libdevice.10.bc
Checking library... ok
Include directory configuration variable:
CUDA_INCLUDE_PATH=/usr/local/cuda/include
Finding include directory from NVIDIA NVCC Wheel
Located at /local/home/pnklein/programming/nonbrazil/myenv/lib/python3.12/site-packages/nvidia/cuda_runtime/include
Checking include directory... ok
\__NumPy Information_\_
NumPy Version : 2.3.4
NumPy Supported SIMD features : ('MMX', 'SSE', 'SSE2', 'SSE3', 'SSSE3', 'SSE41', 'POPCNT', 'SSE42', 'AVX', 'F16C', 'FMA3', 'AVX2')
NumPy Supported SIMD dispatch : ('SSSE3', 'SSE41', 'POPCNT', 'SSE42', 'AVX', 'F16C', 'FMA3', 'AVX2', 'AVX512F', 'AVX512CD', 'AVX512_KNL', 'AVX512_KNM', 'AVX512_SKX', 'AVX512_CNL')
NumPy Supported SIMD baseline : ('SSE', 'SSE2', 'SSE3')
NumPy AVX512_SKX support detected : False
\__SVML Information_\_
SVML State, config.USING_SVML : False
SVML Library Loaded : False
llvmlite Using SVML Patched LLVM : False
SVML Operational : False
\__Threading Layer Information_\_
TBB Threading Layer Available : False
+--> Disabled due to Unknown import problem.
OpenMP Threading Layer Available : True
+-->Vendor: GNU
Workqueue Threading Layer Available : True
+-->Workqueue imported successfully.
\__Numba Environment Variable Information_\_
None found.
\__Conda Information_\_
Conda not available.
\__Installed Packages_\_
Package Version
\------------------------ -----------
airium 0.2.7
annotated-types 0.7.0
boto3 1.40.74
botocore 1.40.74
cuda-bindings 12.9.4
cuda-cccl 0.3.4
cuda-core 0.4.2
cuda-pathfinder 1.3.2
cuda-python 12.9.4
cuda-toolkit 12.9.1
cupy-cuda12x 13.6.0
Cython 3.2.1
docstring_parser 0.17.0
duckdb 1.4.2
fastrlock 0.8.3
iniconfig 2.3.0
jmespath 1.0.1
llvmlite 0.45.1
mypy_extensions 1.1.0
narwhals 2.11.0
numba 0.62.1
numba-cuda 0.20.1
numpy 2.3.4
nvidia-cuda-cccl-cu12 12.9.27
nvidia-cuda-nvcc-cu12 12.9.86
nvidia-cuda-nvrtc-cu12 12.9.86
nvidia-cuda-runtime-cu12 12.9.79
nvidia-nvjitlink-cu12 12.9.86
packaging 25.0
pandas 2.3.3
pip 25.3
plotly 6.4.0
pluggy 1.6.0
psutil 7.1.3
pyarrow 20.0.0
pydantic 2.12.4
pydantic_core 2.41.5
Pygments 2.19.2
pytest 9.0.1
python-dateutil 2.9.0.post0
pytz 2025.2
PyYAML 6.0.3
s3transfer 0.14.0
schema 0.7.8
setuptools 80.9.0
six 1.17.0
typed-argument-parser 1.11.0
typing_extensions 4.15.0
typing-inspect 0.9.0
typing-inspection 0.4.2
tzdata 2025.2
urllib3 2.5.0
No errors reported.
\__Warning log_\_
Warning: Conda not available.
Error was \[Errno 2\] No such file or directory: 'conda'
I can’t see anything wrong with the versions in the numba -s output you have provided - is that from the working machine?
pnk
November 29, 2025, 2:58am
5
That is from the machine where things are not working.