Sylpheed/tools/ppc-manual/fpu/fcmpo.md

fcmpo — Floating Compare Ordered

Category: Floating-Point · Form: X · Opcode: 0xfc000040

Assembler Mnemonics

Mnemonic	XML entry	Flags	Description
fcmpo	fcmpo	—	Floating Compare Ordered

Syntax

fcmpo [CRFD], [FA], [FB]

Encoding

fcmpo — form X

• Opcode word: 0xfc000040
• Primary opcode (bits 0–5): 63
• Extended opcode: 32
• Synchronising: no

Bits	Field	Meaning
0–5	OPCD	primary opcode
6–10	RT/FRT/VRT	destination
11–15	RA/FRA/VRA	source A
16–20	RB/FRB/VRB	source B
21–30	XO	extended opcode (10 bits)
31	Rc	record-form flag

Operands

Field	Role	Description
FA	fcmpo: read	Source A floating-point register (fr0–fr31).
FB	fcmpo: read	Source B floating-point register.
CRFD	fcmpo: write	CR destination field (crf, 0–7).
FPSCR	fcmpo: write	Floating-Point Status and Control Register.

Register Effects

fcmpo

• Reads (always): FA, FB
• Reads (conditional): none
• Writes (always): CRFD, FPSCR
• Writes (conditional): none

Status-Register Effects

• fcmpo: FPSCR updated per IEEE-754 flags (FX, FEX, FPRF, FR, FI, exceptions).

Operation (pseudocode)

; No hand-written pseudocode for this instruction yet.
; The authoritative semantics are the Canary emitter snapshot under
; Implementation References; about half of Canary's emitters open
; with the PPC-style definition as a comment (`RD <- (RA) + (RB)`).
; Every side effect is also enumerated in the Register Effects and
; Status-Register Effects tables above.

C Translation Example

/* No hand-written C yet. Translate the Canary emitter snapshot   */
/* under Implementation References; its HIR maps directly:        */
/*   f.LoadGPR(n) / f.StoreGPR(n, v)  -> r[n] / r[n] = v          */
/*   f.LoadFPR / StoreFPR, f.LoadVR / StoreVR -> f[n], v[n]        */
/*   f.Load(ea, T), f.Store(ea, v) -> raw read / write; emitters   */
/*     wrap them in f.ByteSwap for the big-endian guest value      */
/*   f.UpdateCR(n, v)  -> CR field n from v's LOW 32 BITS vs 0     */
/*   f.LoadCA / f.StoreCA -> xer.CA;  f.StoreSAT -> vscr.SAT       */
/*   i.XO.RA, i.D.DS, ... -> the bit-fields listed under Operands  */
/* The Register Effects and Status-Register Effects tables above  */
/* enumerate every side effect a faithful translation must emit.  */

Implementation References

fcmpo

• Canary XML: tools/ppc-instructions.xml — search for mnem="fcmpo"
• Canary emitter: src/xenia/cpu/ppc/ppc_emit_fpu.cc:362
• Sylpheed opcode: crates/sylpheed-ppc/src/opcode.rs:68
• Sylpheed decoder: crates/sylpheed-ppc/src/decoder.rs:1016

Canary emitter (frozen snapshot @ f21ebd49e9)

int InstrEmit_fcmpo(PPCHIRBuilder& f, const InstrData& i) {
  return InstrEmit_fcmpx_(f, i, true);
}

// ── delegates to (src/xenia/cpu/ppc/ppc_emit_fpu.cc:329) ──
int InstrEmit_fcmpx_(PPCHIRBuilder& f, const InstrData& i, bool ordered) {
  // if (FRA) is a NaN or (FRB) is a NaN then
  //   c <- 0b0001
  // else if (FRA) < (FRB) then
  //   c <- 0b1000
  // else if (FRA) > (FRB) then
  //   c <- 0b0100
  // else {
  //   c <- 0b0010
  // }
  // FPCC <- c
  // CR[4*BF:4*BF+3] <- c
  // if (FRA) is an SNaN or (FRB) is an SNaN then
  //   VXSNAN <- 1

  // TODO(benvanik): update FPCC for mffsx/etc
  // TODO(benvanik): update VXSNAN
  const uint32_t crf = i.X.RT >> 2;
  Value* ra = f.LoadFPR(i.X.RA);
  Value* rb = f.LoadFPR(i.X.RB);

  Value* nan = f.Or(f.IsNan(ra), f.IsNan(rb));
  f.StoreContext(offsetof(PPCContext, cr0) + (4 * crf) + 3, nan);
  Value* not_nan = f.Xor(nan, f.LoadConstantInt8(0x01));

  Value* lt = f.And(not_nan, f.CompareSLT(ra, rb));
  f.StoreContext(offsetof(PPCContext, cr0) + (4 * crf) + 0, lt);
  Value* gt = f.And(not_nan, f.CompareSGT(ra, rb));
  f.StoreContext(offsetof(PPCContext, cr0) + (4 * crf) + 1, gt);
  Value* eq = f.And(not_nan, f.CompareEQ(ra, rb));
  f.StoreContext(offsetof(PPCContext, cr0) + (4 * crf) + 2, eq);
  return 0;
}

Special Cases & Edge Conditions

• Ordered compare. Same CR-field semantics as fcmpu (LT/GT/EQ/SO), but NaN inputs raise additional FPSCR exceptions:
  • Either operand NaN → FPSCR[VXVC] = 1 (invalid-operation: compare on QNaN/SNaN).
  • Either operand signalling NaN → also FPSCR[VXSNAN] = 1.
  • All NaN cases also set FX = 1 and VX = 1.
• Canary's fcmpo is fcmpu. Both call InstrEmit_fcmpx_, which ignores its ordered parameter: the FPSCR exception bits (VXVC, VXSNAN) and FPCC are not modelled. Title code that polls FPSCR for compare-class invalid-operation will not observe it.
• CR field bits.
  • LT (bit 0) — FRA < FRB
  • GT (bit 1) — FRA > FRB
  • EQ (bit 2) — FRA == FRB
  • SO (bit 3) — unordered (NaN involved)
• +0 and -0 compare equal.
• No Rc bit.
• FPSCR side effects. Hardware updates FPSCR[FPCC], FX, VX, and (on NaN) VXVC/VXSNAN. Canary only updates the CR field.
• Use case. Ordered compares are required by C/C++ semantics for <, >, <=, >= (which must signal on NaN per IEEE-754). fcmpu corresponds to the C ==/!= semantics (which do not signal).
• Encoding. X-form, primary 63, XO 32.

Related Instructions

• fcmpux — unordered compare; identical CR result, no VXVC.
• mcrf, mcrfs, mfcr — fan-out CR fields after compare.
• bc, bclr, bcctr — conditional branches consume LT/GT/EQ/SO.
• fselx — branch-free alternative for single-key compares.
• mcrfs, mffsx — move FPSCR/CR.

IBM Reference

• AIX 7.3 — fcmpo (Floating Compare Ordered)
• PowerISA v2.07B, Book I, Chapter 4 — Floating-Point Processor (fcmpo raises VXVC on QNaN; both raise VXSNAN on SNaN).