# `fsub.s`

[RISC-V Documentation](https://asm-editor.specy.app/documentation/risc-v.md) › [Instructions](https://asm-editor.specy.app/documentation/risc-v/instruction.md)

Web page: https://asm-editor.specy.app/documentation/risc-v/instruction/fsub.s

**Floating subtract**

Category: Floating point

Subtracts the third register from the second in single precision and stores the difference in the first.

[RISC-V manual: fsub.s](https://docs.riscv.org/reference/isa/v20260120/unpriv/f-st-ext.html#norm:fsub-s_op)

Used in: [Floating point and the F and D extensions](https://asm-editor.specy.app/learn/courses/risc-v/think-in-assembly/floating-point-and-the-f-and-d-extensions.md) (RISC-V assembly)

## Forms

- `fsub.s f1, f2, f3, rm`

## Pseudo-instructions

The processor does not have these forms. The assembler replaces each of them with one or more real instructions. Open a row to see them.

- `fsub.s f1, f2, f3`
  - `fsub.s f1, f2, f3` → `fsub.s f1, f2, f3, dyn`

## Operands

- `f1`: Any floating-point register, `f0` to `f31`, or its named alias, such as `ft0`, `fa0` or `ft11`.
- `f2`: A floating-point register. It may be the same register as `f1`.
- `f3`: A floating-point register. It may be the same register as `f1`.
- `rm`: A rounding mode: `rne` (nearest, ties to even), `rtz` (toward zero), `rdn` (toward negative infinity), `rup` (toward positive infinity), `rmm` (nearest, ties away from zero) or `dyn` (use the rounding mode in `fcsr`).

## Example

```riscv
.data
a: .float 6.0
b: .float 2.0
.text
la t0, a
flw f1, 0(t0)
la t1, b
flw f2, 0(t1)
# f1 (ft1)=6 and f2 (ft2)=2; fsub.s writes 4.
fsub.s f3, f1, f2
```
