We updated our CUDA version and Nvidia driver:
Now when we try to run a GPU job in cryoSPARC, it stalls when starting the pipeline and never begins running. There seems to still be a CUDA error occurring. For one, cryoSPARC cannot ‘see’ our GPUs:
(base) [cryosparc@C05195 ~]$ /home/cryosparc/cryosparc_worker/bin/cryosparcw gpulist
Traceback (most recent call last):
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 258, in ensure_initialized
self.cuInit(0)
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 352, in safe_cuda_api_call
return self._check_cuda_python_error(fname, libfn(*args))
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 412, in _check_cuda_python_error
raise CudaAPIError(retcode, msg)
numba.cuda.cudadrv.driver.CudaAPIError: [CUresult.CUDA_ERROR_UNKNOWN] Call to cuInit results in CUDA_ERROR_UNKNOWN
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/home/cryosparc/cryosparc_worker/bin/connect.py", line 28, in print_gpu_list
num_devs = len(cuda.gpus)
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/devices.py", line 49, in __len__
return len(self.lst)
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/devices.py", line 26, in __getattr__
numdev = driver.get_device_count()
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 425, in get_device_count
return self.cuDeviceGetCount()
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 296, in __getattr__
self.ensure_initialized()
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 262, in ensure_initialized
raise CudaSupportError(f"Error at driver init: {description}")
numba.cuda.cudadrv.error.CudaSupportError: Error at driver init: Call to cuInit results in CUDA_ERROR_UNKNOWN (999)
Additionally, when we try to connect the cryoSPARC worker to our nodes, the CPUs connect fine but a very similar CUDA error occurs when it tries to connect to the GPUs:
(base) [cryosparc@C05195 cryosparc_worker]$ bin/cryosparcw connect \
> --worker localhost \
> --master localhost \
> --port 39000 \
> --update \
> --gpus 0,1,2,3
---------------------------------------------------------------
CRYOSPARC CONNECT --------------------------------------------
---------------------------------------------------------------
Attempting to register worker localhost to command localhost:39002
Connecting as unix user cryosparc
Will register using ssh string: cryosparc@localhost
If this is incorrect, you should re-run this command with the flag --sshstr <ssh string>
---------------------------------------------------------------
Connected to master.
---------------------------------------------------------------
Current connected workers:
localhost
---------------------------------------------------------------
Worker will be registered with 48 CPUs.
---------------------------------------------------------------
Updating target localhost
Current configuration:
cache_path : /data
cache_quota_mb : None
cache_reserve_mb : 10000
desc : None
gpus : [{'id': 0, 'mem': 11554717696, 'name': 'NVIDIA GeForce RTX 2080 Ti'}, {'id': 1, 'mem': 11554717696, 'name': 'NVIDIA GeForce RTX 2080 Ti'}, {'id': 2, 'mem': 11554717696, 'name': 'NVIDIA GeForce RTX 2080 Ti'}, {'id': 3, 'mem': 11554324480, 'name': 'NVIDIA GeForce RTX 2080 Ti'}]
hostname : localhost
lane : default
monitor_port : None
name : localhost
resource_fixed : {'SSD': True}
resource_slots : {'CPU': [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, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47], 'GPU': [0, 1, 2, 3], 'RAM': [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, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47]}
ssh_str : cryosparc@localhost
title : Worker node localhost
type : node
worker_bin_path : /home/cryosparc/cryosparc_worker/bin/cryosparcw
---------------------------------------------------------------
Autodetecting available GPUs...
Traceback (most recent call last):
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 258, in ensure_initialized
self.cuInit(0)
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 352, in safe_cuda_api_call
return self._check_cuda_python_error(fname, libfn(*args))
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 412, in _check_cuda_python_error
raise CudaAPIError(retcode, msg)
numba.cuda.cudadrv.driver.CudaAPIError: [CUresult.CUDA_ERROR_UNKNOWN] Call to cuInit results in CUDA_ERROR_UNKNOWN
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "bin/connect.py", line 165, in <module>
gpu_devidxs = check_gpus()
File "bin/connect.py", line 101, in check_gpus
num_devs = print_gpu_list()
File "bin/connect.py", line 28, in print_gpu_list
num_devs = len(cuda.gpus)
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/devices.py", line 49, in __len__
return len(self.lst)
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/devices.py", line 26, in __getattr__
numdev = driver.get_device_count()
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 425, in get_device_count
return self.cuDeviceGetCount()
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 296, in __getattr__
self.ensure_initialized()
File "/home/cryosparc/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/numba/cuda/cudadrv/driver.py", line 262, in ensure_initialized
raise CudaSupportError(f"Error at driver init: {description}")
numba.cuda.cudadrv.error.CudaSupportError: Error at driver init: Call to cuInit results in CUDA_ERROR_UNKNOWN (999)
We’re wondering if there is some incompatibility or lack of cross-talk between the CUDA version we updated and the CUDA toolkit that comes pre-installed with v4.4. Should we uninstall and reinstall cryoSPARC and import our existing projects, or is there a simpler potential solution? Thanks!