Failed building wheel for pycuda

...In file included from src/cpp/cuda.cpp:4:src/cpp/cuda.hpp:14:10: fatal error: cuda.h: 没有那个文件或目录14 | #include <cuda.h>|          ^~~~~~~~compilation terminated.error: command '/usr/bin/g++' failed with exit code 1[end of output]note: This error originates from a subprocess, and is likely not a problem with pip.ERROR: Failed building wheel for pycudaDEPRECATION: Building 'nvidia-pyindex' using the legacy setup.py bdist_wheel mechanism, which will be removed in a future version. pip 25.3 will enforce this behaviour change. A possible replacement is to use the standardized build interface by setting the `--use-pep517` option, (possibly combined with `--no-build-isolation`), or adding a `pyproject.toml` file to the source tree of 'nvidia-pyindex'. Discussion can be found at https://github.com/pypa/pip/issues/6334Building wheel for nvidia-pyindex (setup.py) ... doneCreated wheel for nvidia-pyindex: filename=nvidia_pyindex-1.0.9-py3-none-any.whl size=8510 sha256=ddc833868c516fe46860ba0ce38fd6cfea94912dc8107825be013a422e53e973Stored in directory: /home/wangguisen/.cache/pip/wheels/30/9f/d7/10a73e63166091b744d10c631596f2c7e53ac14a41b9246de5
Successfully built nvidia-pyindex
Failed to build pycuda
ERROR: Failed to build installable wheels for some pyproject.toml based projects (pycuda)

这个错误表明在编译 pycuda 包时，系统无法找到 cuda.h 文件。这通常是因为以下几个原因：

1. CUDA Toolkit 未安装：cuda.h 文件是 CUDA Toolkit 的一部分，如果没有安装 CUDA Toolkit，就会出现这个错误。
2. CUDA Toolkit 安装路径不正确：即使安装了 CUDA Toolkit，如果其路径未正确配置，编译器也可能找不到 cuda.h 文件。
3. 环境变量未设置或设置不正确：编译器需要通过环境变量（如 CUDA_HOME 和 LD_LIBRARY_PATH）来找到 CUDA Toolkit 的路径。

先运行：

nvcc -V

如果提示：

找不到命令 “nvcc”，但可以通过以下软件包安装它：
apt install nvidia-cuda-toolkit
请联系您的管理员。

先确定是不是安装了，很有可能是安装了但是环境变量没有配：

ls /usr/local/cuda# 如果你看到类似以下的输出，说明 CUDA Toolkit 已经安装：
# bin  include  lib64  share  ... 

如果你已经安装了 CUDA Toolkit，但系统无法找到 nvcc 命令，可能是因为 CUDA Toolkit 的路径没有添加到系统的 PATH 环境变量中。你可以通过以下命令临时添加路径：

export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

为了永久生效，你可以将这些行添加到你的 ~/.bashrc 文件中：

echo 'export PATH=/usr/local/cuda/bin:$PATH' >> ~/.bashrc
echo 'export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH' >> ~/.bashrc
source ~/.bashrc

nvcc -V

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2024 NVIDIA Corporation
Built on Thu_Mar_28_02:18:24_PDT_2024
Cuda compilation tools, release 12.4, V12.4.131
Build cuda_12.4.r12.4/compiler.34097967_0

如果你没有安装 CUDA Toolkit，可以根据提示安装，注意和cuda版本对应，记得写入环境变量。

sudo apt-get install nvidia-cuda-toolkit