ValueError: Unknown CUDA arch (8.6) or GPU not supported

[wutianze]

When trying to install pylot in my base system, I came up with the following error.
/home/sauron/.local/lib/python3.8/site-packages/torch/include/ATen/core/TensorBody.h:322:7: note: declared here 322 | T * data() const { | ^~~~ Traceback (most recent call last): File "setup.py", line 4, in <module> setup( File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/__init__.py", line 87, in setup return distutils.core.setup(**attrs) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/core.py", line 148, in setup return run_commands(dist) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/core.py", line 163, in run_commands dist.run_commands() File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/dist.py", line 967, in run_commands self.run_command(cmd) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/dist.py", line 1214, in run_command super().run_command(command) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/dist.py", line 986, in run_command cmd_obj.run() File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/command/build.py", line 136, in run self.run_command(cmd_name) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/cmd.py", line 313, in run_command self.distribution.run_command(command) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/dist.py", line 1214, in run_command super().run_command(command) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/dist.py", line 986, in run_command cmd_obj.run() File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/command/build_ext.py", line 79, in run _build_ext.run(self) File "/home/sauron/.local/lib/python3.8/site-packages/Cython/Distutils/old_build_ext.py", line 186, in run _build_ext.build_ext.run(self) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/command/build_ext.py", line 339, in run self.build_extensions() File "/home/sauron/.local/lib/python3.8/site-packages/torch/utils/cpp_extension.py", line 372, in build_extensions build_ext.build_extensions(self) File "/home/sauron/.local/lib/python3.8/site-packages/Cython/Distutils/old_build_ext.py", line 195, in build_extensions _build_ext.build_ext.build_extensions(self) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/command/build_ext.py", line 448, in build_extensions self._build_extensions_serial() File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/command/build_ext.py", line 473, in _build_extensions_serial self.build_extension(ext) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/command/build_ext.py", line 202, in build_extension _build_ext.build_extension(self, ext) File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/command/build_ext.py", line 528, in build_extension objects = self.compiler.compile(sources, File "/home/sauron/.local/lib/python3.8/site-packages/setuptools/_distutils/ccompiler.py", line 574, in compile self._compile(obj, src, ext, cc_args, extra_postargs, pp_opts) File "/home/sauron/.local/lib/python3.8/site-packages/torch/utils/cpp_extension.py", line 288, in unix_wrap_compile "'-fPIC'"] + cflags + _get_cuda_arch_flags(cflags) File "/home/sauron/.local/lib/python3.8/site-packages/torch/utils/cpp_extension.py", line 1027, in _get_cuda_arch_flags raise ValueError("Unknown CUDA arch ({}) or GPU not supported".format(arch)) ValueError: Unknown CUDA arch (8.6) or GPU not supported [x] Downloading the CARLA 0.9.10.1 simulator...
My GPU is rtx3090ti, and the cuda version is 11.6

[wutianze]

The problem is:
The AnyNet which pylot uses only supports cuda10, while the 30 series cards use cuda > 11

[pschafhalter]

Does this problem also occur when running Pylot using the docker container?

If you don't need to use Pylot with AnyNet, you could also try commenting out the following lines of the installation script and re-running:

pylot/install.sh

Lines 100 to 104 in 765bbfa

	###### Download AnyNet depth estimation code #####
	echo "[x] Cloning the AnyNet depth estimation code..."
	cd $PYLOT_HOME/dependencies/
	git clone https://github.com/mileyan/AnyNet.git
	cd AnyNet/models/spn_t1/ ; python3 setup.py clean ; python3 setup.py build

[wutianze]

Does this problem also occur when running Pylot using the docker container?

If you don't need to use Pylot with AnyNet, you could also try commenting out the following lines of the installation script and re-running:

pylot/install.sh

Lines 100 to 104 in 765bbfa

	###### Download AnyNet depth estimation code #####
	echo "[x] Cloning the AnyNet depth estimation code..."
	cd $PYLOT_HOME/dependencies/
	git clone https://github.com/mileyan/AnyNet.git
	cd AnyNet/models/spn_t1/ ; python3 setup.py clean ; python3 setup.py build

Thanks for your reply. This problem would not occur when running docker container since the container simulates a cuda 10 environment.
And I will try commenting out AnyNet, thanks again. By the way, I wonder if it is possible to use UE4 to make some changes to the pylot scenario(the city staff) based on your containers instead of using the base system.

[pschafhalter]

Pylot is bundled with the CARLA simulator. We provide some flags for configuring the simulator (e.g. # of vehicles and pedestrians, weather, which town is used -- run python3 pylot.py --helpfull for details), but you may need modify CARLA directly depending on your requirements.

[wutianze]

Pylot is bundled with the CARLA simulator. We provide some flags for configuring the simulator (e.g. # of vehicles and pedestrians, weather, which town is used -- run python3 pylot.py --helpfull for details), but you may need modify CARLA directly depending on your requirements.

Yes, I will try modifying CARLA in the container you provided, no worrying about building a CARLA environment again, hope that will work.