xenia-rs/migration/project-root/ppc-manual/vmx/vsubfp.md

# `vsubfp` — Vector Subtract Floating Point

> **Category:** [VMX (Altivec)](../categories/vmx.md) · **Form:** [VX](../forms/VX.md) · **Opcode:** `0x1000004a`

<!-- GENERATED: BEGIN -->

## Assembler Mnemonics

| Mnemonic | XML entry | Flags | Description |
| --- | --- | --- | --- |
| `vsubfp` | `vsubfp` | — | Vector Subtract Floating Point |
| `vsubfp128` | `vsubfp128` | — | Vector128 Subtract Floating Point |

## Syntax

```asm
vsubfp [VD], [VA], [VB]
vsubfp128 [VD], [VA], [VB]
```

## Encoding

### `vsubfp` — form `VX`

- **Opcode word:** `0x1000004a`
- **Primary opcode (bits 0–5):** `4`
- **Extended opcode:** `74`
- **Synchronising:** no

| Bits | Field | Meaning |
| --- | --- | --- |
| 0–5 | `OPCD` | primary opcode (4) |
| 6–10 | `VRT/VD` | destination vector register |
| 11–15 | `VRA/VA` | source A vector register |
| 16–20 | `VRB/VB` | source B vector register |
| 21–31 | `XO` | extended opcode (11 bits) |

### `vsubfp128` — form `VX128`

- **Opcode word:** `0x14000050`
- **Primary opcode (bits 0–5):** `5`
- **Extended opcode:** `80`
- **Synchronising:** no

| Bits | Field | Meaning |
| --- | --- | --- |
| 0–5 | `OPCD` | primary opcode (4 or 5) |
| 6–10 | `VD128l` | destination low 5 bits |
| 11–15 | `VA128l` | source A low 5 bits |
| 16–20 | `VB128l` | source B low 5 bits |
| 21 | `VA128H` | source A high bit |
| 22 | `—` | reserved |
| 23–25 | `VC` | optional VC / XO sub-field |
| 26 | `VA128h` | source A middle bit |
| 27 | `—` | reserved |
| 28–29 | `VD128h` | destination high 2 bits |
| 30–31 | `VB128h` | source B high 2 bits |

## Operands

| Field | Role | Description |
| --- | --- | --- |
| `VA` | vsubfp: read; vsubfp128: read | Source A vector register. |
| `VB` | vsubfp: read; vsubfp128: read | Source B vector register. |
| `VD` | vsubfp: write; vsubfp128: write | Destination vector register. |

## Register Effects

### `vsubfp`

- **Reads (always):** `VA`, `VB`
- **Reads (conditional):** _none_
- **Writes (always):** `VD`
- **Writes (conditional):** _none_

### `vsubfp128`

- **Reads (always):** `VA`, `VB`
- **Reads (conditional):** _none_
- **Writes (always):** `VD`
- **Writes (conditional):** _none_

## Status-Register Effects

_No condition-register or status-register effects._

## Operation (pseudocode)

```
for each 32-bit float lane i in 0..3:
    VD[i] <- VA[i] − VB[i]
```

## C Translation Example

```c
/* C translation: the xenia-rs interpreter arm below in           */
/* Implementation References is the authoritative semantic        */
/* snapshot. Translate it line-by-line:                            */
/*   - ctx.gpr[N]  -> r[N]       (or f[]/v[] for FPRs/VRs)        */
/*   - mem.read_u*/write_u* -> mem_read_u*_be / mem_write_u*_be   */
/*   - ctx.update_cr_signed(fld, v) -> update_cr_signed(fld, v)   */
/*   - ctx.xer_ca / xer_ov / xer_so -> xer.CA / xer.OV / xer.SO   */
/* The Register Effects and Status-Register Effects tables above  */
/* enumerate every side effect a faithful translation must emit.  */
```

## Implementation References

**`vsubfp`**
- xenia-canary XML: [`tools/ppc-instructions.xml` — search for `mnem="vsubfp"`](../../xenia-canary/tools/ppc-instructions.xml)
- xenia-canary emit: [`src/xenia/cpu/ppc/ppc_emit_altivec.cc:1686`](../../xenia-canary/src/xenia/cpu/ppc/ppc_emit_altivec.cc#L1686)
- xenia-rs opcode: [`crates/xenia-cpu/src/opcode.rs:125`](../../xenia-rs/crates/xenia-cpu/src/opcode.rs#L125)
- xenia-rs decoder: [`crates/xenia-cpu/src/decoder.rs:445`](../../xenia-rs/crates/xenia-cpu/src/decoder.rs#L445)
- xenia-rs interpreter: [`crates/xenia-cpu/src/interpreter.rs:2012-2024`](../../xenia-rs/crates/xenia-cpu/src/interpreter.rs#L2012-L2024)
<details><summary>xenia-rs interpreter body (frozen snapshot)</summary>

```rust
        PpcOpcode::vsubfp => {
            // PPCBUG-435.
            let a = ctx.vr[instr.ra()].as_f32x4();
            let b = ctx.vr[instr.rb()].as_f32x4();
            let mut r = [0f32; 4];
            for i in 0..4 {
                let ai = vmx::flush_denorm(a[i]);
                let bi = vmx::flush_denorm(b[i]);
                r[i] = vmx::flush_denorm(ai - bi);
            }
            ctx.vr[instr.rd()] = xenia_types::Vec128::from_f32x4_array(r);
            ctx.pc += 4;
        }
```
</details>

**`vsubfp128`**
- xenia-canary XML: [`tools/ppc-instructions.xml` — search for `mnem="vsubfp128"`](../../xenia-canary/tools/ppc-instructions.xml)
- xenia-canary emit: [`src/xenia/cpu/ppc/ppc_emit_altivec.cc:1689`](../../xenia-canary/src/xenia/cpu/ppc/ppc_emit_altivec.cc#L1689)
- xenia-rs opcode: [`crates/xenia-cpu/src/opcode.rs:125`](../../xenia-rs/crates/xenia-cpu/src/opcode.rs#L125)
- xenia-rs decoder: [`crates/xenia-cpu/src/decoder.rs:611`](../../xenia-rs/crates/xenia-cpu/src/decoder.rs#L611)
- xenia-rs interpreter: [`crates/xenia-cpu/src/interpreter.rs:2025-2037`](../../xenia-rs/crates/xenia-cpu/src/interpreter.rs#L2025-L2037)
<details><summary>xenia-rs interpreter body (frozen snapshot)</summary>

```rust
        PpcOpcode::vsubfp128 => {
            // PPCBUG-435.
            let a = ctx.vr[instr.va128()].as_f32x4();
            let b = ctx.vr[instr.vb128()].as_f32x4();
            let mut r = [0f32; 4];
            for i in 0..4 {
                let ai = vmx::flush_denorm(a[i]);
                let bi = vmx::flush_denorm(b[i]);
                r[i] = vmx::flush_denorm(ai - bi);
            }
            ctx.vr[instr.vd128()] = xenia_types::Vec128::from_f32x4_array(r);
            ctx.pc += 4;
        }
```
</details>

<!-- GENERATED: END -->

## Special Cases & Edge Conditions

- **Lane-wise IEEE-754 binary32 subtract.** Each of the four lanes computes `VD[i] = VA[i] − VB[i]`, rounded to nearest.
- **`VSCR[NJ]` honoured.** Denormals flushed to zero when `NJ = 1` (the Xenon boot default).
- **NaN propagation.** A NaN in either operand propagates to the destination lane.
- **`±∞ − ±∞` → NaN.** No exception, no VSCR[SAT] set.
- **No FPSCR update.** VMX float ops are independent of the scalar FPU's status register.
- **Big-endian lane indexing.**
- **VMX128 sibling [`vsubfp128`](vsubfp128.md).**
- **Aliasing legal.** `vsubfp v3, v3, v4` is fine.

## Related Instructions

- [`vaddfp`](vaddfp.md) — lane-wise float add.
- [`vmaddfp`](vmaddfp.md), [`vnmsubfp`](vnmsubfp.md) — fused multiply-accumulate variants.
- [`vmaxfp`](vmaxfp.md), [`vminfp`](vminfp.md) — IEEE-754-aware max/min.
- [`vcmpeqfp`](vcmpeqfp.md), [`vcmpgtfp`](vcmpgtfp.md), [`vcmpgefp`](vcmpgefp.md), [`vcmpbfp`](vcmpbfp.md) — compares.

## IBM Reference

- [AIX 7.3 — `vsubfp` (Vector Subtract Floating Point)](https://www.ibm.com/docs/en/aix/7.3.0?topic=set-vsubfp-vector-subtract-floating-point-instruction)
- [IBM AltiVec Technology Programmer's Interface Manual, Chapter 5 — Floating-Point Arithmetic](https://www.nxp.com/docs/en/reference-manual/ALTIVECPIM.pdf)