一文详解如何用GPU来运行Python代码/基于Python自制一个文件解压缩小工具

前几天捣鼓了一下Ubuntu，正是想用一下我旧电脑上的N卡，可以用GPU来跑代码，体验一下多核的快乐，感兴趣的小伙伴快跟随小编一起了解一下吧

简介

前几天捣鼓了一下Ubuntu，正是想用一下我旧电脑上的N卡，可以用GPU来跑代码，体验一下多核的快乐。

还好我这破电脑也是支持Cuda的：

1 2 3 4 5 6 7 8 9 10 11 12 13

$ sudo lshw -C display   *-display                        description: 3D controller        product: GK208M [GeForce GT 740M]        vendor: NVIDIA Corporation        physical id: 0        bus info: pci@0000:01:00.0        version: a1        width: 64 bits        clock: 33MHz        capabilities: pm msi pciexpress bus_master cap_list rom        configuration: driver=nouveau latency=0        resources: irq:35 memory:f0000000-f0ffffff memory:c0000000-cfffffff memory:d0000000-d1ffffff ioport:6000(size=128)

安装相关工具

首先安装一下Cuda的开发工具，命令如下：

1	$ sudo apt install nvidia-cuda-toolkit

查看一下相关信息：

1 2 3 4 5 6

$ nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2021 NVIDIA Corporation Built on Thu_Nov_18_09:45:30_PST_2021 Cuda compilation tools, release 11.5, V11.5.119 Build cuda_11.5.r11.5/compiler.30672275_0

通过Conda安装相关的依赖包：

1	conda install numba & conda install cudatoolkit

通过pip安装也可以，一样的。

测试与驱动安装

简单测试了一下，发觉报错了：

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

$ /home/larry/anaconda3/bin/python /home/larry/code/pkslow-samples/python/src/main/python/cuda/test1.py Traceback (most recent call last):   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/cudadrv/driver.py", line 246, in ensure_initialized     self.cuInit(0)   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/cudadrv/driver.py", line 319, in safe_cuda_api_call     self._check_ctypes_error(fname, retcode)   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/cudadrv/driver.py", line 387, in _check_ctypes_error     raise CudaAPIError(retcode, msg) numba.cuda.cudadrv.driver.CudaAPIError: [100] Call to cuInit results in CUDA_ERROR_NO_DEVICE During handling of the above exception, another exception occurred: Traceback (most recent call last):   File "/home/larry/code/pkslow-samples/python/src/main/python/cuda/test1.py", line 15, in <module>     gpu_print[1, 2]()   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/compiler.py", line 862, in __getitem__     return self.configure(*args)   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/compiler.py", line 857, in configure     return _KernelConfiguration(self, griddim, blockdim, stream, sharedmem)   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/compiler.py", line 718, in __init__     ctx = get_context()   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/cudadrv/devices.py", line 220, in get_context     return _runtime.get_or_create_context(devnum)   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/cudadrv/devices.py", line 138, in get_or_create_context     return self._get_or_create_context_uncached(devnum)   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/cudadrv/devices.py", line 153, in _get_or_create_context_uncached     with driver.get_active_context() as ac:   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/cudadrv/driver.py", line 487, in __enter__     driver.cuCtxGetCurrent(byref(hctx))   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/cudadrv/driver.py", line 284, in __getattr__     self.ensure_initialized()   File "/home/larry/anaconda3/lib/python3.9/site-packages/numba/cuda/cudadrv/driver.py", line 250, 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_NO_DEVICE (100)

网上搜了一下，发现是驱动问题。通过Ubuntu自带的工具安装显卡驱动：

 

还是失败：

1 2	$ nvidia-smi NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

最后，通过命令行安装驱动，成功解决这个问题：

1	$ sudo apt install nvidia-driver-470

检查后发现正常了：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

$ nvidia-smi Wed Dec  7 22:13:49 2022       +-----------------------------------------------------------------------------+ | NVIDIA-SMI 470.161.03   Driver Version: 470.161.03   CUDA Version: 11.4     | |-------------------------------+-----