RuntimeError: CUDA is not available
mahdian-organization opened this issue · comments
a.m commented
Hi there,
I got an error when I ran my code. My graphics card is a GeForce GT 730M
.
Warp CUDA error: Failed to get function cuGetProcAddress
Warp 0.10.1 initialized:
CUDA devices not available
Devices:
"cpu" | Intel64 Family 6 Model 60 Stepping 3, GenuineIntel
Kernel cache: C:\Users\a.mahdianrad\AppData\Local\NVIDIA Corporation\warp\Cache\0.10.1
Warp CUDA error: Function cuCtxGetCurrent_f: a suitable driver entry point was not found
Warp CUDA error 36 (W:\a9ae500d09a78409\warp\native\warp.cu:155)
Traceback (most recent call last):
File "E:\PycharmProjects\testSimulation\main.py", line 902, in <module>
initScene()
File "E:\PycharmProjects\testSimulation\main.py", line 567, in initScene
cloth = Cloth(clothY, clothNumX, clothNumY, clothSpacing, sphereCenter, sphereRadius)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "E:\PycharmProjects\testSimulation\main.py", line 115, in __init__
self.pos = wp.array(pos, dtype=wp.vec3, device="cuda")
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "E:\PycharmProjects\testSimulation\venv\Lib\site-packages\warp\types.py", line 1273, in __init__
self._init_from_data(data, dtype, shape, device, copy, pinned)
File "E:\PycharmProjects\testSimulation\venv\Lib\site-packages\warp\types.py", line 1396, in _init_from_data
device = warp.get_device(device)
^^^^^^^^^^^^^^^^^^^^^^^
File "E:\PycharmProjects\testSimulation\venv\Lib\site-packages\warp\context.py", line 2500, in get_device
return runtime.get_device(ident)
^^^^^^^^^^^^^^^^^^^^^^^^^
File "E:\PycharmProjects\testSimulation\venv\Lib\site-packages\warp\context.py", line 2323, in get_device
return self.get_current_cuda_device()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "E:\PycharmProjects\testSimulation\venv\Lib\site-packages\warp\context.py", line 2354, in get_current_cuda_device
raise RuntimeError("CUDA is not available")
RuntimeError: CUDA is not available
Nicolas Capens commented
Sorry about the unclear error message. The GeForce GT 730M is unfortunately not supported by Warp. We require CUDA Compute Capability 5.2 as a minimum, and the GT 730M only supports 3.0. Warp uses CUDA Tookit 11.5 or higher, which no longer supports GPUs with Compute Capability 3.0.
We'll improve our documentation and the error message in a future release.
Nicolas Capens commented
d43b12c improves the error reporting and documentation.