blob: 68098b822d66e742bc1cc7e10f580ef5f9dddfed [file] [log] [blame] [raw]
<!DOCTYPE html>
<html>
<head>
<meta charset="UTF-8">
<link href="style.css" type="text/css" rel="stylesheet">
<title>VFNMSUB132PS/VFNMSUB213PS/VFNMSUB231PS — Fused Negative Multiply-Subtract of Packed Single-Precision Floating-Point Values </title></head>
<body>
<h1>VFNMSUB132PS/VFNMSUB213PS/VFNMSUB231PS — Fused Negative Multiply-Subtract of Packed Single-Precision Floating-Point Values</h1>
<table>
<tr>
<th>Opcode/Instruction</th>
<th>Op/En</th>
<th>64/32 -bit Mode</th>
<th>CPUID Feature Flag</th>
<th>Description</th></tr>
<tr>
<td>
<p>VEX.DDS.128.66.0F38.W0 9E /r</p>
<p>VFNMSUB132PS <em>xmm0, xmm1, xmm2/m128</em></p></td>
<td>A</td>
<td>V/V</td>
<td>FMA</td>
<td>Multiply packed single-precision floating-point values from <em>xmm0</em> and <em>xmm2/mem</em>, negate the multiplica-tion result and subtract <em>xmm1</em> and put result in <em>xmm0</em>.</td></tr>
<tr>
<td>
<p>VEX.DDS.128.66.0F38.W0 AE /r</p>
<p>VFNMSUB213PS <em>xmm0, xmm1, xmm2/m128</em></p></td>
<td>A</td>
<td>V/V</td>
<td>FMA</td>
<td>Multiply packed single-precision floating-point values from <em>xmm0</em> and <em>xmm1</em>, negate the multiplication result and subtract <em>xmm2/mem</em> and put result in <em>xmm0</em>.</td></tr>
<tr>
<td>
<p>VEX.DDS.128.66.0F38.W0 BE /r</p>
<p>VFNMSUB231PS <em>xmm0, xmm1, xmm2/m128</em></p></td>
<td>A</td>
<td>V/V</td>
<td>FMA</td>
<td>Multiply packed single-precision floating-point values from <em>xmm1</em> and <em>xmm2/mem</em>, negate the multiplica-tion result and subtract <em>xmm0</em> and put result in <em>xmm0</em>.</td></tr>
<tr>
<td>
<p>VEX.DDS.256.66.0F38.W0 9E /r</p>
<p>VFNMSUB132PS <em>ymm0, ymm1, ymm2/m256</em></p></td>
<td>A</td>
<td>V/V</td>
<td>FMA</td>
<td>Multiply packed single-precision floating-point values from <em>ymm0</em> and <em>ymm2/mem</em>, negate the multiplica-tion result and subtract <em>ymm1</em> and put result in <em>ymm0</em>.</td></tr>
<tr>
<td>
<p>VEX.DDS.256.66.0F38.W0 AE /r</p>
<p>VFNMSUB213PS <em>ymm0, ymm1, ymm2/m256</em></p></td>
<td>A</td>
<td>V/V</td>
<td>FMA</td>
<td>Multiply packed single-precision floating-point values from <em>ymm0</em> and <em>ymm1</em>, negate the multiplication result and subtract <em>ymm2/mem</em> and put result in <em>ymm0</em>.</td></tr>
<tr>
<td>
<p>VEX.DDS.256.66.0F38.0 BE /r</p>
<p>VFNMSUB231PS <em>ymm0, ymm1, ymm2/m256</em></p></td>
<td>A</td>
<td>V/V</td>
<td>FMA</td>
<td>Multiply packed single-precision floating-point values from <em>ymm1</em> and <em>ymm2/mem,</em> negate the multiplica-tion result and subtract <em>ymm0</em> and put result in <em>ymm0</em>.</td></tr></table>
<h3>Instruction Operand Encoding</h3>
<table>
<tr>
<td>Op/En</td>
<td>Operand 1</td>
<td>Operand 2</td>
<td>Operand 3</td>
<td>Operand 4</td></tr>
<tr>
<td>A</td>
<td>ModRM:reg (r, w)</td>
<td>VEX.vvvv (r)</td>
<td>ModRM:r/m (r)</td>
<td>NA</td></tr></table>
<h2>Description</h2>
<p>VFNMSUB132PS: Multiplies the four or eight packed single-precision floating-point values from the first source operand to the four or eight packed single-precision floating-point values in the third source operand. From negated infinite precision intermediate results, subtracts the four or eight packed single-precision floating-point values in the second source operand, performs rounding and stores the resulting four or eight packed single-preci-sion floating-point values to the destination operand (first source operand).</p>
<p>VFNMSUB213PS: Multiplies the four or eight packed single-precision floating-point values from the second source operand to the four or eight packed single-precision floating-point values in the first source operand. From negated infinite precision intermediate results, subtracts the four or eight packed single-precision floating-point values in the third source operand, performs rounding and stores the resulting four or eight packed single-precision floating-point values to the destination operand (first source operand).</p>
<p>VFNMSUB231PS: Multiplies the four or eight packed single-precision floating-point values from the second source to the four or eight packed single-precision floating-point values in the third source operand. From negated infinite precision intermediate results, subtracts the four or eight packed single-precision floating-point values in the first source operand, performs rounding and stores the resulting four or eight packed single-precision floating-point values to the destination operand (first source operand).</p>
<p>VEX.128 encoded version: The destination operand (also first source operand) is a XMM register and encoded in reg_field. The second source operand is a XMM register and encoded in VEX.vvvv. The third source operand is a</p>
<p>XMM register or a 128-bit memory location and encoded in rm_field. The upper 128 bits of the YMM destination register are zeroed.</p>
<p>VEX.256 encoded version: The destination operand (also first source operand) is a YMM register and encoded in reg_field. The second source operand is a YMM register and encoded in VEX.vvvv. The third source operand is a YMM register or a 256-bit memory location and encoded in rm_field.</p>
<p>Compiler tools may optionally support a complementary mnemonic for each instruction mnemonic listed in the opcode/instruction column of the summary table. The behavior of the complementary mnemonic in situations involving NANs are governed by the definition of the instruction mnemonic defined in the opcode/instruction column. See also Section 14.5.1, “FMA Instruction Operand Order and Arithmetic Behavior” in the <em>Intel® 64 and IA-32 Architectures Software Developer’s Manual, Volume 1</em>.</p>
<h2>Operation</h2>
<p>In the operations below, "+", "-", and "*" symbols represent addition, subtraction, and multiplication operations with infinite precision inputs and outputs (no rounding).</p>
<p><strong>VFNMSUB132PS DEST, SRC2, SRC3</strong></p>
<pre>IF (VEX.128) THEN
MAXVL =4
ELSEIF (VEX.256)
MAXVL = 8
FI
For i = 0 to MAXVL-1 {
n = 32*i;
DEST[n+31:n] ← RoundFPControl_MXCSR( - (DEST[n+31:n]*SRC3[n+31:n]) - SRC2[n+31:n])
}
IF (VEX.128) THEN
DEST[VLMAX-1:128] ← 0
FI</pre>
<p><strong>VFNMSUB213PS DEST, SRC2, SRC3</strong></p>
<pre>IF (VEX.128) THEN
MAXVL =4
ELSEIF (VEX.256)
MAXVL = 8
FI
For i = 0 to MAXVL-1 {
n = 32*i;
DEST[n+31:n] ← RoundFPControl_MXCSR( - (SRC2[n+31:n]*DEST[n+31:n]) - SRC3[n+31:n])
}
IF (VEX.128) THEN
DEST[VLMAX-1:128] ← 0
FI</pre>
<p><strong>VFNMSUB231PS DEST, SRC2, SRC3</strong></p>
<pre>IF (VEX.128) THEN
MAXVL =4
ELSEIF (VEX.256)
MAXVL = 8
FI
For i = 0 to MAXVL-1 {
n = 32*i;
DEST[n+31:n] ← RoundFPControl_MXCSR( - (SRC2[n+31:n]*SRC3[n+31:n]) - DEST[n+31:n])
}
IF (VEX.128) THEN
DEST[VLMAX-1:128] ← 0
FI</pre>
<h2>Intel C/C++ Compiler Intrinsic Equivalent</h2>
<p>VFNMSUB132PS: __m128 _mm_fnmsub_ps (__m128 a, __m128 b, __m128 c);</p>
<p>VFNMSUB213PS: __m128 _mm_fnmsub_ps (__m128 a, __m128 b, __m128 c);</p>
<p>VFNMSUB231PS: __m128 _mm_fnmsub_ps (__m128 a, __m128 b, __m128 c);</p>
<p>VFNMSUB132PS: __m256 _mm256_fnmsub_ps (__m256 a, __m256 b, __m256 c);</p>
<p>VFNMSUB213PS: __m256 _mm256_fnmsub_ps (__m256 a, __m256 b, __m256 c);</p>
<p>VFNMSUB231PS: __m256 _mm256_fnmsub_ps (__m256 a, __m256 b, __m256 c);</p>
<h2>SIMD Floating-Point Exceptions</h2>
<p>Overflow, Underflow, Invalid, Precision, Denormal</p>
<h2>Other Exceptions</h2>
<p>See Exceptions Type 2</p></body></html>