ValueError: Unknown CUDA arch (8.6) or GPU not supported 解决RTX30 运行深度学习代码报错

将conda环境所在文件夹中的cpp_extension.py内容从：

    named_arches = collections.OrderedDict([
        ('Kepler+Tesla', '3.7'),
        ('Kepler', '3.5+PTX'),
        ('Maxwell+Tegra', '5.3'),
        ('Maxwell', '5.0;5.2+PTX'),
        ('Pascal', '6.0;6.1+PTX'),
        ('Volta', '7.0+PTX'),
        ('Turing', '7.5+PTX'),
    ])

    supported_arches = ['3.5', '3.7', '5.0', '5.2', '5.3', '6.0', '6.1', '6.2',
                        '7.0', '7.2', '7.5']

更改为：

    named_arches = collections.OrderedDict([
        ('Kepler+Tesla', '3.7'),
        ('Kepler', '3.5+PTX'),
        ('Maxwell+Tegra', '5.3'),
        ('Maxwell', '5.0;5.2+PTX'),
        ('Pascal', '6.0;6.1+PTX'),
        ('Volta', '7.0+PTX'),
        ('Turing', '7.5+PTX'),
        ('Ampere', '8.0;8.6+PTX'),
    ])

    supported_arches = ['3.5', '3.7', '5.0', '5.2', '5.3', '6.0', '6.1', '6.2',
                        '7.0', '7.2', '7.5', '8.0', '8.6']

记录个没想到的情况，运行某算法时，因为设备是RTX30，查到说是只能CUDA11+。
但一次conda create时直接覆盖了一个CUDA11.2的环境，并在这个环境上安装成功了cuda10和pytorch1.3

——————————————————————————————————————
后记：
然而，接下来又报错
RuntimeError: cublas runtime error : the GPU program failed to execute at /opt/conda/conda-bld/pytorch_1573049310284/work/aten/src/THC/THCBlas.cu:331

经查询是cuda版本10.0不能再RTX30系列上跑，更换版本吧。。没招