Add support for devices with compute capability < 6.0

[tobbez]

Without this change, building for devices with compute capability < 6.0 fails with:

quant_cuda_kernel.cu(149): error: no instance of overloaded function "atomicAdd" matches the argument list
            argument types are: (double *, double)
          detected during instantiation of "void VecQuant2MatMulKernel(const scalar_t *, const int *, scalar_t *, const scalar_t *, const scalar_t *, int, int, int, int) [with scalar_t=double]"
(87): here