103 lines
5 KiB
HTML
103 lines
5 KiB
HTML
<!DOCTYPE html>
|
||
<html xmlns="http://www.w3.org/1999/xhtml" xmlns:svg="http://www.w3.org/2000/svg" xmlns:x86="http://www.felixcloutier.com/x86"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8"><link rel="stylesheet" type="text/css" href="style.css"></link><title>VSCALEFSD
|
||
— Scale Scalar Float64 Values With Float64 Values</title></head><body><header><nav><ul><li><a href='index.html'>Index</a></li><li>December 2023</li></ul></nav></header><h1>VSCALEFSD
|
||
— Scale Scalar Float64 Values With Float64 Values</h1>
|
||
|
||
<table>
|
||
<tr>
|
||
<th>Opcode/Instruction</th>
|
||
<th>Op/En</th>
|
||
<th>64/32 bit Mode Support</th>
|
||
<th>CPUID Feature Flag</th>
|
||
<th>Description</th></tr>
|
||
<tr>
|
||
<td>EVEX.LLIG.66.0F38.W1 2D /r VSCALEFSD xmm1 {k1}{z}, xmm2, xmm3/m64{er}</td>
|
||
<td>A</td>
|
||
<td>V/V</td>
|
||
<td>AVX512F</td>
|
||
<td>Scale the scalar double precision floating-point values in xmm2 using the value from xmm3/m64. Under writemask k1.</td></tr></table>
|
||
<h2 id="instruction-operand-encoding">Instruction Operand Encoding<a class="anchor" href="#instruction-operand-encoding">
|
||
¶
|
||
</a></h2>
|
||
<table>
|
||
<tr>
|
||
<th>Op/En</th>
|
||
<th>Tuple Type</th>
|
||
<th>Operand 1</th>
|
||
<th>Operand 2</th>
|
||
<th>Operand 3</th>
|
||
<th>Operand 4</th></tr>
|
||
<tr>
|
||
<td>A</td>
|
||
<td>Tuple1 Scalar</td>
|
||
<td>ModRM:reg (w)</td>
|
||
<td>EVEX.vvvv (r)</td>
|
||
<td>ModRM:r/m (r)</td>
|
||
<td>N/A</td></tr></table>
|
||
<h3 id="description">Description<a class="anchor" href="#description">
|
||
¶
|
||
</a></h3>
|
||
<p>Performs a floating-point scale of the scalar double precision floating-point value in the first source operand by multiplying it by 2 to the power of the double precision floating-point value in second source operand.</p>
|
||
<p>The equation of this operation is given by:</p>
|
||
<p>xmm1 := xmm2*2<sup>floor(xmm3)</sup>.</p>
|
||
<p>Floor(xmm3) means maximum integer value ≤ xmm3.</p>
|
||
<p>If the result cannot be represented in double precision, then the proper overflow response (for positive scaling operand), or the proper underflow response (for negative scaling operand) is issued. The overflow and underflow responses are dependent on the rounding mode (for IEEE-compliant rounding), as well as on other settings in MXCSR (exception mask bits, FTZ bit), and on the SAE bit.</p>
|
||
<p>EVEX encoded version: The first source operand is an XMM register. The second source operand is an XMM register or a memory location. The destination operand is an XMM register conditionally updated with writemask k1.</p>
|
||
<p>Handling of special-case input values are listed in <a href='vscalefpd.html#tbl-5-39'>Table 5-39</a> and <a href='vscalefpd.html#tbl-5-40'>Table 5-40</a>.</p>
|
||
<h3 id="operation">Operation<a class="anchor" href="#operation">
|
||
¶
|
||
</a></h3>
|
||
<pre>SCALE(SRC1, SRC2)
|
||
{
|
||
; Check for denormal operands
|
||
TMP_SRC2 := SRC2
|
||
TMP_SRC1 := SRC1
|
||
IF (SRC2 is denormal AND MXCSR.DAZ) THEN TMP_SRC2=0
|
||
IF (SRC1 is denormal AND MXCSR.DAZ) THEN TMP_SRC1=0
|
||
/* SRC2 is a 64 bits floating-point value */
|
||
DEST[63:0] := TMP_SRC1[63:0] * POW(2, Floor(TMP_SRC2[63:0]))
|
||
}
|
||
</pre>
|
||
<h4 id="vscalefsd--evex-encoded-version-">VSCALEFSD (EVEX encoded version)<a class="anchor" href="#vscalefsd--evex-encoded-version-">
|
||
¶
|
||
</a></h4>
|
||
<pre>IF (EVEX.b= 1) and SRC2 *is a register*
|
||
THEN
|
||
SET_ROUNDING_MODE_FOR_THIS_INSTRUCTION(EVEX.RC);
|
||
ELSE
|
||
SET_ROUNDING_MODE_FOR_THIS_INSTRUCTION(MXCSR.RC);
|
||
FI;
|
||
IF k1[0] OR *no writemask*
|
||
THEN DEST[63:0] := SCALE(SRC1[63:0], SRC2[63:0])
|
||
ELSE
|
||
IF *merging-masking* ; merging-masking
|
||
THEN *DEST[63:0] remains unchanged*
|
||
ELSE ; zeroing-masking
|
||
DEST[63:0] := 0
|
||
FI
|
||
FI;
|
||
DEST[127:64] := SRC1[127:64]
|
||
DEST[MAXVL-1:128] := 0
|
||
</pre>
|
||
<h3 id="intel-c-c++-compiler-intrinsic-equivalent">Intel C/C++ Compiler Intrinsic Equivalent<a class="anchor" href="#intel-c-c++-compiler-intrinsic-equivalent">
|
||
¶
|
||
</a></h3>
|
||
<pre>VSCALEFSD __m128d _mm_scalef_round_sd(__m128d a, __m128d b, int);
|
||
</pre>
|
||
<pre>VSCALEFSD __m128d _mm_mask_scalef_round_sd(__m128d s, __mmask8 k, __m128d a, __m128d b, int);
|
||
</pre>
|
||
<pre>VSCALEFSD __m128d _mm_maskz_scalef_round_sd(__mmask8 k, __m128d a, __m128d b, int);
|
||
</pre>
|
||
<h3 class="exceptions" id="simd-floating-point-exceptions">SIMD Floating-Point Exceptions<a class="anchor" href="#simd-floating-point-exceptions">
|
||
¶
|
||
</a></h3>
|
||
<p>Overflow, Underflow, Invalid, Precision, Denormal (for Src1).</p>
|
||
<p>Denormal is not reported for Src2.</p>
|
||
<h3 class="exceptions" id="other-exceptions">Other Exceptions<a class="anchor" href="#other-exceptions">
|
||
¶
|
||
</a></h3>
|
||
<p>See <span class="not-imported">Table 2-47</span>, “Type E3 Class Exception Conditions.”</p><footer><p>
|
||
This UNOFFICIAL, mechanically-separated, non-verified reference is provided for convenience, but it may be
|
||
inc<span style="opacity: 0.2">omp</span>lete or b<sub>r</sub>oke<sub>n</sub> in various obvious or non-obvious
|
||
ways. Refer to <a href="https://software.intel.com/en-us/download/intel-64-and-ia-32-architectures-sdm-combined-volumes-1-2a-2b-2c-2d-3a-3b-3c-3d-and-4">Intel® 64 and IA-32 Architectures Software Developer’s Manual</a> for anything serious.
|
||
</p></footer></body></html>
|