Is it possible to use it directly in the Linux kernel code?

I want to take advantage of the GPU features(Parallel processing) within the Linux kernel

For example, I want to use it in the kernel network stack, such as the ip_rcv() function.

I wonder if this is possible.

I would appreciate it if you let me know.