fsw

The `fsw` instruction stores a single-precision floating-point value in `fs2` to memory at address `xs1` + `imm`. It does not modify the bits being transferred; in particular, the payloads of non-canonical NaNs are preserved.

fsw fs2, imm(xs1)