Linux kernel does not support floating-point operations during compilation

The project I am currently working on aims to save energy by modifying the kernel task scheduling strategy, which requires operating on the Linux kernel. Recently, an algorithm involved floating-point operations, and an error message was generated during the compilation of the Linux kernel, showing that the SSE register is not supported. The screenshot is shown below. After some understanding, it was found that it is related to gcc disabling sse and sse2, involving software, floating-point, hard floating-point, FPU, and other related content. I hope to obtain an exact modification method to ensure that the compilation passes smoothly, and that the compiled kernel can be opened and used smoothly (because the relevant information (- mno sse, - mno mmx, etc.) was deleted in the Makefile before, the compilation was very smooth, but the compiled kernel cannot be opened)