// -*- C -*-

// Simulator definition for the MIPS DSP REV 2 ASE.
// Copyright (C) 2007-2024 Free Software Foundation, Inc.
// Contributed by MIPS Technologies, Inc.
// Written by Chao-ying Fu (fu@mips.com).
//
// This file is part of the MIPS sim
//
// This program is free software; you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation; either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program.  If not, see <http://www.gnu.org/licenses/>.


// op: 0 = ADD, 1 = SUB
// sat: 0 = no saturation, 1 = saturation
:function:::void:do_u_ph_op:int rd, int rs, int rt, int op, int sat
{
  int i;
  uint32_t h0;
  uint16_t h1, h2;
  uint32_t v1 = GPR[rs];
  uint32_t v2 = GPR[rt];
  uint32_t result = 0;
  for (i = 0; i < 32; i += 16, v1 >>= 16, v2 >>= 16)
    {
      h1 = (uint16_t)(v1 & 0xffff);
      h2 = (uint16_t)(v2 & 0xffff);
      if (op == 0) // ADD
	h0 = (uint32_t)h1 + (uint32_t)h2;
      else // SUB
	h0 = (uint32_t)h1 - (uint32_t)h2;
      if (op == 0 && (h0 > (uint32_t)0x0000ffff)) // ADD SAT
	{
	  DSPCR |= DSPCR_OUFLAG4;
	  if (sat == 1)
	    h0 = 0xffff;
	}
      else if (op == 1 && h1 < h2) // SUB SAT
	{
	  DSPCR |= DSPCR_OUFLAG4;
	  if (sat == 1)
	    h0 = 0x0;
	}
      result |= ((uint32_t)((uint16_t)h0) << i);
    }
  GPR[rd] = EXTEND32 (result);
}

// op: 0 = ADD, 1 = SUB
// round: 0 = no rounding, 1 = rounding
:function:::void:do_uh_qb_op:int rd, int rs, int rt, int op, int round
{
  int i;
  uint32_t h0;
  uint8_t h1, h2;
  uint32_t v1 = GPR[rs];
  uint32_t v2 = GPR[rt];
  uint32_t result = 0;
  for (i = 0; i < 32; i += 8, v1 >>= 8, v2 >>= 8)
    {
      h1 = (uint8_t)(v1 & 0xff);
      h2 = (uint8_t)(v2 & 0xff);
      if (op == 0) // ADD
	h0 = (uint32_t)h1 + (uint32_t)h2;
      else // SUB
	h0 = (uint32_t)h1 - (uint32_t)h2;
      if (round == 1)
	h0 = (h0 + 1) >> 1;
      else
	h0 = h0 >> 1;
      result |= ((uint32_t)((uint8_t)h0) << i);
    }
  GPR[rd] = EXTEND32 (result);
}

// op: 0 = EQ, 1 = LT, 2 = LE
:function:::void:do_qb_cmpgdu:int rd, int rs, int rt, int op
{
  int i, j;
  uint32_t v1 = GPR[rs];
  uint32_t v2 = GPR[rt];
  uint8_t h1, h2;
  uint32_t result = 0;
  uint32_t mask;
  for (i = 0, j = 0; i < 32; i += 8, j++, v1 >>= 8, v2 >>= 8)
    {
      h1 = (uint8_t)(v1 & 0xff);
      h2 = (uint8_t)(v2 & 0xff);
      mask = ~(1 << (DSPCR_CCOND_SHIFT + j));
      DSPCR &= mask;
      if (op == 0) // EQ
	{
	  result |= ((h1 == h2) << j);
	  DSPCR |= ((h1 == h2) << (DSPCR_CCOND_SHIFT + j));
	}
      else if (op == 1) // LT
	{
	  result |= ((h1 < h2) << j);
	  DSPCR |= ((h1 < h2) << (DSPCR_CCOND_SHIFT + j));
	}
      else // LE
	{
	  result |= ((h1 <= h2) << j);
	  DSPCR |= ((h1 <= h2) << (DSPCR_CCOND_SHIFT + j));
	}
    }
  GPR[rd] = EXTEND32 (result);
}

// op: 0 = DPA 1 = DPS
:function:::void:do_w_ph_dot_product:int ac, int rs, int rt, int op
{
  int i;
  uint32_t v1 = GPR[rs];
  uint32_t v2 = GPR[rt];
  int16_t h1, h2;
  int32_t result;
  uint32_t lo = DSPLO(ac);
  uint32_t hi = DSPHI(ac);
  int64_t prod = (int64_t)((((uint64_t)hi) << 32) + (uint64_t)lo);
  for (i = 0; i < 32; i += 16, v1 >>= 16, v2 >>= 16)
    {
      h1 = (int16_t)(v1 & 0xffff);
      h2 = (int16_t)(v2 & 0xffff);
      result = (int32_t)h1 * (int32_t)h2;
      if (op == 0) // DPA
        prod += (int64_t)result;
      else // DPS
        prod -= (int64_t)result;
    }
  DSPLO(ac) = EXTEND32 (prod);
  DSPHI(ac) = EXTEND32 (prod >> 32);
}

// round: 0 = no rounding, 1 = rounding
:function:::void:do_w_mulq:int rd, int rs, int rt, int round
{
  uint32_t v1 = GPR[rs];
  uint32_t v2 = GPR[rt];
  int32_t w1, w2;
  int64_t prod;
  uint32_t result;
  w1 = (int32_t) v1;
  w2 = (int32_t) v2;
  if (w1 == (int32_t) 0x80000000 && w2 == (int32_t) 0x80000000)
    {
      DSPCR |= DSPCR_OUFLAG5;
      prod = 0x7fffffff;
    }
  else
    {
      prod = ((int64_t) w1 * (int64_t) w2) << 1;
      if (round == 1)
	prod += 0x0000000080000000LL;
      prod = prod >> 32;
    }
  result = (uint32_t) prod;
  GPR[rd] = EXTEND32 (result);
}

// round: 0 = no rounding, 1 = rounding
:function:::void:do_precr_sra:int rt, int rs, int sa, int round
{
  uint32_t v1 = GPR[rt];
  uint32_t v2 = GPR[rs];
  int32_t w1 = (int32_t) v1;
  int32_t w2 = (int32_t) v2;
  int32_t result;
  if (sa != 0)
    {
      if (round == 1 && (w1 & (1 << (sa - 1))))
	w1 = (w1 >> sa) + 1;
      else
	w1 = w1 >> sa;

      if (round == 1 && (w2 & (1 << (sa - 1))))
	w2 = (w2 >> sa) + 1;
      else
	w2 = w2 >> sa;
    }
  result = (w1 << 16) | (w2 & 0xffff);
  GPR[rt] = EXTEND32 (result);
}

// round: 0 = no rounding, 1 = rounding
:function:::void:do_qb_shra:int rd, int rt, int shift, int round
{
  int i;
  int8_t q0;
  uint32_t v1 = GPR[rt];
  uint32_t result = 0;
  for (i = 0; i < 32; i += 8, v1 >>= 8)
    {
      q0 = (int8_t)(v1 & 0xff);
      if (shift != 0)
 	{
	  if (round == 1 && (q0 & (1 << (shift - 1))))
	    q0 = (q0 >> shift) + 1;
	  else
	    q0 = q0 >> shift;
 	}
      result |= ((uint32_t)((uint8_t)q0) << i);
    }
  GPR[rd] = EXTEND32 (result);
}

:function:::void:do_ph_shrl:int rd, int rt, int shift
{
  int i;
  uint16_t h0;
  uint32_t v1 = GPR[rt];
  uint32_t result = 0;
  for (i = 0; i < 32; i += 16, v1 >>= 16)
    {
      h0 = (uint16_t)(v1 & 0xffff);
      h0 = h0 >> shift;
      result |= ((uint32_t)h0 << i);
    }
  GPR[rd] = EXTEND32 (result);
}

// op: 0 = ADD, 1 = SUB
// round: 0 = no rounding, 1 = rounding
:function:::void:do_qh_ph_op:int rd, int rs, int rt, int op, int round
{
  int i;
  int32_t h0;
  int16_t h1, h2;
  uint32_t v1 = GPR[rs];
  uint32_t v2 = GPR[rt];
  uint32_t result = 0;
  for (i = 0; i < 32; i += 16, v1 >>= 16, v2 >>= 16)
    {
      h1 = (int16_t)(v1 & 0xffff);
      h2 = (int16_t)(v2 & 0xffff);
      if (op == 0) // ADD
	h0 = (int32_t)h1 + (int32_t)h2;
      else // SUB
	h0 = (int32_t)h1 - (int32_t)h2;
      if (round == 1)
	h0 = (h0 + 1) >> 1;
      else
	h0 = h0 >> 1;
      result |= ((uint32_t)((uint16_t)h0) << i);
    }
  GPR[rd] = EXTEND32 (result);
}

// op: 0 = ADD, 1 = SUB
// round: 0 = no rounding, 1 = rounding
:function:::void:do_qh_w_op:int rd, int rs, int rt, int op, int round
{
  int64_t v0;
  int32_t v1 = (int32_t)GPR[rs];
  int32_t v2 = (int32_t)GPR[rt];
  if (op == 0) // ADD
    v0 = (int64_t)v1 + (int64_t)v2;
  else // SUB
    v0 = (int64_t)v1 - (int64_t)v2;
  if (round == 1)
    v0 = (v0 + 1) >> 1;
  else
    v0 = v0 >> 1;
  GPR[rd] = EXTEND32 (v0);
}

// op: 0 = DPAX, 1 = DPSX
:function:::void:do_x_w_ph_dot_product:int ac, int rs, int rt, int op
{
  int i;
  uint32_t v1 = GPR[rs];
  uint32_t v2 = GPR[rt];
  int16_t h1, h2;
  int32_t result;
  uint32_t lo = DSPLO(ac);
  uint32_t hi = DSPHI(ac);
  int64_t prod = (int64_t)((((uint64_t)hi) << 32) + (uint64_t)lo);
  for (i = 0; i < 32; i += 16, v1 >>= 16, v2 <<= 16)
    {
      h1 = (int16_t)(v1 & 0xffff);
      h2 = (int16_t)((v2 & 0xffff0000) >> 16);
      result = (int32_t)h1 * (int32_t)h2;
      if (op == 0) // DPAX
        prod += (int64_t)result;
      else // DPSX
        prod -= (int64_t)result;
    }
  DSPLO(ac) = EXTEND32 (prod);
  DSPHI(ac) = EXTEND32 (prod >> 32);
}

// op: 0 = DPAQX, 1 = DPSQX
// sat: 0 = no saturation, 1 = saturation of the accumulator
:function:::void:do_qx_w_ph_dot_product:int ac, int rs, int rt, int op, int sat
{
  int i;
  uint32_t v1 = GPR[rs];
  uint32_t v2 = GPR[rt];
  int16_t h1, h2;
  int32_t result;
  uint32_t lo = DSPLO(ac);
  uint32_t hi = DSPHI(ac);
  int64_t prod = (int64_t)((((uint64_t)hi) << 32) + (uint64_t)lo);
  int64_t max, min;
  for (i = 0; i < 32; i += 16, v1 >>= 16, v2 <<= 16)
    {
      h1 = (int16_t)(v1 & 0xffff);
      h2 = (int16_t)((v2 & 0xffff0000) >> 16);
      if (h1 == (int16_t)0x8000 && h2 == (int16_t)0x8000)
	{
	  DSPCR |= (1 << (DSPCR_OUFLAG_SHIFT + ac));
	  result = 0x7fffffff;
	}
      else
	result = ((int32_t)h1 * (int32_t)h2) << 1;
      if (op == 0) // DPAQX
        prod += (int64_t)result;
      else // DPSQX
        prod -= (int64_t)result;
    }
  // Saturation on the accumulator.
  if (sat == 1)
    {
      max = (int64_t) 0x7fffffffLL;
      min = (int64_t) 0xffffffff80000000LL;
      if (prod > max)
	{
	  DSPCR |= (1 << (DSPCR_OUFLAG_SHIFT + ac));
	  prod = max;
	}
      else if (prod < min)
	{
	  DSPCR |= (1 << (DSPCR_OUFLAG_SHIFT + ac));
	  prod = min;
	}
    }
  DSPLO(ac) = EXTEND32 (prod);
  DSPHI(ac) = EXTEND32 (prod >> 32);
}

011111,00000,5.RT,5.RD,00001,010010:SPECIAL3:32::ABSQ_S.QB
"absq_s.qb r<RD>, r<RT>"
*dsp2:
{
  do_qb_s_absq (SD_, RD, RT);
}

011111,5.RS,5.RT,5.RD,01000,010000:SPECIAL3:32::ADDU.PH
"addu.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_u_ph_op (SD_, RD, RS, RT, 0, 0);
}

011111,5.RS,5.RT,5.RD,01100,010000:SPECIAL3:32::ADDU_S.PH
"addu_s.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_u_ph_op (SD_, RD, RS, RT, 0, 1);
}

011111,5.RS,5.RT,5.RD,00000,011000:SPECIAL3:32::ADDUH.QB
"adduh.qb r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_uh_qb_op (SD_, RD, RS, RT, 0, 0);
}

011111,5.RS,5.RT,5.RD,00010,011000:SPECIAL3:32::ADDUH_R.QB
"adduh_r.qb r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_uh_qb_op (SD_, RD, RS, RT, 0, 1);
}

011111,5.RS,5.RT,5.SA,00000,110001:SPECIAL3:32::APPEND
"append r<RT>, r<RS>, <SA>"
*dsp2:
{
  do_append (SD_, RT, RS, SA);
}

011111,5.RS,5.RT,000,2.BP,10000,110001:SPECIAL3:32::BALIGN
"balign r<RT>, r<RS>, <BP>"
*dsp2:
{
  do_balign (SD_, RT, RS, BP);
}

011111,5.RS,5.RT,5.RD,11000,010001:SPECIAL3:32::CMPGDU.EQ.QB
"cmpgdu.eq.qb r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qb_cmpgdu (SD_, RD, RS, RT, 0);
}

011111,5.RS,5.RT,5.RD,11001,010001:SPECIAL3:32::CMPGDU.LT.QB
"cmpgdu.lt.qb r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qb_cmpgdu (SD_, RD, RS, RT, 1);
}

011111,5.RS,5.RT,5.RD,11010,010001:SPECIAL3:32::CMPGDU.LE.QB
"cmpgdu.le.qb r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qb_cmpgdu (SD_, RD, RS, RT, 2);
}

011111,5.RS,5.RT,000,2.AC,00000,110000:SPECIAL3:32::DPA.W.PH
"dpa.w.ph ac<AC>, r<RS>, r<RT>"
*dsp2:
{
  do_w_ph_dot_product (SD_, AC, RS, RT, 0);
}

011111,5.RS,5.RT,000,2.AC,00001,110000:SPECIAL3:32::DPS.W.PH
"dps.w.ph ac<AC>, r<RS>, r<RT>"
*dsp2:
{
  do_w_ph_dot_product (SD_, AC, RS, RT, 1);
}

011111,5.RS,5.RT,5.RD,01100,011000:SPECIAL3:32::MUL.PH
"mul.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_ph_op (SD_, RD, RS, RT, 2, 0);
}

011111,5.RS,5.RT,5.RD,01110,011000:SPECIAL3:32::MUL_S.PH
"mul_s.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_ph_op (SD_, RD, RS, RT, 2, 1);
}

011111,5.RS,5.RT,5.RD,10111,011000:SPECIAL3:32::MULQ_RS.W
"mulq_rs.w r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_w_mulq (SD_, RD, RS, RT, 1);
}

011111,5.RS,5.RT,5.RD,11110,010000:SPECIAL3:32::MULQ_S.PH
"mulq_s.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_ph_mulq (SD_, RD, RS, RT, 0);
}

011111,5.RS,5.RT,5.RD,10110,011000:SPECIAL3:32::MULQ_S.W
"mulq_s.w r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_w_mulq (SD_, RD, RS, RT, 0);
}

011111,5.RS,5.RT,000,2.AC,00010,110000:SPECIAL3:32::MULSA.W.PH
"mulsa.w.ph ac<AC>, r<RS>, r<RT>"
*dsp2:
{
  do_ph_w_mulsa (SD_, AC, RS, RT);
}

011111,5.RS,5.RT,5.RD,01101,010001:SPECIAL3:32::PRECR.QB.PH
"precr.qb.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_ph_qb_precr (SD_, RD, RS, RT);
}

011111,5.RS,5.RT,5.SA,11110,010001:SPECIAL3:32::PRECR_SRA.PH.W
"precr_sra.ph.w r<RT>, r<RS>, <SA>"
*dsp2:
{
  do_precr_sra (SD_, RT, RS, SA, 0);
}

011111,5.RS,5.RT,5.SA,11111,010001:SPECIAL3:32::PRECR_SRA_R.PH.W
"precr_sra_r.ph.w r<RT>, r<RS>, <SA>"
*dsp2:
{
  do_precr_sra (SD_, RT, RS, SA, 1);
}

011111,5.RS,5.RT,5.SA,00001,110001:SPECIAL3:32::PREPEND
"prepend r<RT>, r<RS>, <SA>"
*dsp2:
{
  do_prepend (SD_, RT, RS, SA);
}

011111,00,3.SHIFT3,5.RT,5.RD,00100,010011:SPECIAL3:32::SHRA.QB
"shra.qb r<RD>, r<RT>, <SHIFT3>"
*dsp2:
{
  do_qb_shra (SD_, RD, RT, SHIFT3, 0);
}

011111,00,3.SHIFT3,5.RT,5.RD,00101,010011:SPECIAL3:32::SHRA_R.QB
"shra_r.qb r<RD>, r<RT>, <SHIFT3>"
*dsp2:
{
  do_qb_shra (SD_, RD, RT, SHIFT3, 1);
}

011111,5.RS,5.RT,5.RD,00110,010011:SPECIAL3:32::SHRAV.QB
"shrav.qb r<RD>, r<RT>, r<RS>"
*dsp2:
{
  do_qb_shrav (SD_, RD, RT, RS, 0);
}

011111,5.RS,5.RT,5.RD,00111,010011:SPECIAL3:32::SHRAV_R.QB
"shrav_r.qb r<RD>, r<RT>, r<RS>"
*dsp2:
{
  do_qb_shrav (SD_, RD, RT, RS, 1);
}

011111,0,4.SHIFT4,5.RT,5.RD,11001,010011:SPECIAL3:32::SHRL.PH
"shrl.ph r<RD>, r<RT>, <SHIFT4>"
*dsp2:
{
  do_ph_shrl (SD_, RD, RT, SHIFT4);
}

011111,5.RS,5.RT,5.RD,11011,010011:SPECIAL3:32::SHRLV.PH
"shrlv.ph r<RD>, r<RT>, r<RS>"
*dsp2:
{
  do_ph_shrlv (SD_, RD, RT, RS);
}

011111,5.RS,5.RT,5.RD,01001,010000:SPECIAL3:32::SUBU.PH
"subu.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_u_ph_op (SD_, RD, RS, RT, 1, 0);
}

011111,5.RS,5.RT,5.RD,01101,010000:SPECIAL3:32::SUBU_S.PH
"subu_s.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_u_ph_op (SD_, RD, RS, RT, 1, 1);
}

011111,5.RS,5.RT,5.RD,00001,011000:SPECIAL3:32::SUBUH.QB
"subuh.qb r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_uh_qb_op (SD_, RD, RS, RT, 1, 0);
}

011111,5.RS,5.RT,5.RD,00011,011000:SPECIAL3:32::SUBUH_R.QB
"subuh_r.qb r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_uh_qb_op (SD_, RD, RS, RT, 1, 1);
}

011111,5.RS,5.RT,5.RD,01000,011000:SPECIAL3:32::ADDQH.PH
"addqh.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qh_ph_op (SD_, RD, RS, RT, 0, 0);
}

011111,5.RS,5.RT,5.RD,01010,011000:SPECIAL3:32::ADDQH_R.PH
"addqh_r.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qh_ph_op (SD_, RD, RS, RT, 0, 1);
}

011111,5.RS,5.RT,5.RD,10000,011000:SPECIAL3:32::ADDQH.W
"addqh.w r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qh_w_op (SD_, RD, RS, RT, 0, 0);
}

011111,5.RS,5.RT,5.RD,10010,011000:SPECIAL3:32::ADDQH_R.W
"addqh_r.w r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qh_w_op (SD_, RD, RS, RT, 0, 1);
}

011111,5.RS,5.RT,5.RD,01001,011000:SPECIAL3:32::SUBQH.PH
"subqh.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qh_ph_op (SD_, RD, RS, RT, 1, 0);
}

011111,5.RS,5.RT,5.RD,01011,011000:SPECIAL3:32::SUBQH_R.PH
"subqh_r.ph r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qh_ph_op (SD_, RD, RS, RT, 1, 1);
}

011111,5.RS,5.RT,5.RD,10001,011000:SPECIAL3:32::SUBQH.W
"subqh.w r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qh_w_op (SD_, RD, RS, RT, 1, 0);
}

011111,5.RS,5.RT,5.RD,10011,011000:SPECIAL3:32::SUBQH_R.W
"subqh_r.w r<RD>, r<RS>, r<RT>"
*dsp2:
{
  do_qh_w_op (SD_, RD, RS, RT, 1, 1);
}

011111,5.RS,5.RT,000,2.AC,01000,110000:SPECIAL3:32::DPAX.W.PH
"dpax.w.ph ac<AC>, r<RS>, r<RT>"
*dsp2:
{
  do_x_w_ph_dot_product (SD_, AC, RS, RT, 0);
}

011111,5.RS,5.RT,000,2.AC,01001,110000:SPECIAL3:32::DPSX.W.PH
"dpsx.w.ph ac<AC>, r<RS>, r<RT>"
*dsp2:
{
  do_x_w_ph_dot_product (SD_, AC, RS, RT, 1);
}

011111,5.RS,5.RT,000,2.AC,11000,110000:SPECIAL3:32::DPAQX_S.W.PH
"dpaqx_s.w.ph ac<AC>, r<RS>, r<RT>"
*dsp2:
{
  do_qx_w_ph_dot_product (SD_, AC, RS, RT, 0, 0);
}

011111,5.RS,5.RT,000,2.AC,11010,110000:SPECIAL3:32::DPAQX_SA.W.PH
"dpaqx_sa.w.ph ac<AC>, r<RS>, r<RT>"
*dsp2:
{
  do_qx_w_ph_dot_product (SD_, AC, RS, RT, 0, 1);
}

011111,5.RS,5.RT,000,2.AC,11001,110000:SPECIAL3:32::DPSQX_S.W.PH
"dpsqx_s.w.ph ac<AC>, r<RS>, r<RT>"
*dsp2:
{
  do_qx_w_ph_dot_product (SD_, AC, RS, RT, 1, 0);
}

011111,5.RS,5.RT,000,2.AC,11011,110000:SPECIAL3:32::DPSQX_SA.W.PH
"dpsqx_sa.w.ph ac<AC>, r<RS>, r<RT>"
*dsp2:
{
  do_qx_w_ph_dot_product (SD_, AC, RS, RT, 1, 1);
}