291 lines
12 KiB
HTML
291 lines
12 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>VPMOVWB/VPMOVSWB/VPMOVUSWB
|
|||
|
|
— Down Convert Word to Byte</title></head><body><header><nav><ul><li><a href='index.html'>Index</a></li><li>December 2023</li></ul></nav></header><h1>VPMOVWB/VPMOVSWB/VPMOVUSWB
|
|||
|
|
— Down Convert Word to Byte</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.128.F3.0F38.W0 30 /r VPMOVWB xmm1/m64 {k1}{z}, xmm2</td>
|
|||
|
|
<td>A</td>
|
|||
|
|
<td>V/V</td>
|
|||
|
|
<td>AVX512VL AVX512BW</td>
|
|||
|
|
<td>Converts 8 packed word integers from xmm2 into 8 packed bytes in xmm1/m64 with truncation under writemask k1.</td></tr>
|
|||
|
|
<tr>
|
|||
|
|
<td>EVEX.128.F3.0F38.W0 20 /r VPMOVSWB xmm1/m64 {k1}{z}, xmm2</td>
|
|||
|
|
<td>A</td>
|
|||
|
|
<td>V/V</td>
|
|||
|
|
<td>AVX512VL AVX512BW</td>
|
|||
|
|
<td>Converts 8 packed signed word integers from xmm2 into 8 packed signed bytes in xmm1/m64 using signed saturation under writemask k1.</td></tr>
|
|||
|
|
<tr>
|
|||
|
|
<td>EVEX.128.F3.0F38.W0 10 /r VPMOVUSWB xmm1/m64 {k1}{z}, xmm2</td>
|
|||
|
|
<td>A</td>
|
|||
|
|
<td>V/V</td>
|
|||
|
|
<td>AVX512VL AVX512BW</td>
|
|||
|
|
<td>Converts 8 packed unsigned word integers from xmm2 into 8 packed unsigned bytes in 8mm1/m64 using unsigned saturation under writemask k1.</td></tr>
|
|||
|
|
<tr>
|
|||
|
|
<td>EVEX.256.F3.0F38.W0 30 /r VPMOVWB xmm1/m128 {k1}{z}, ymm2</td>
|
|||
|
|
<td>A</td>
|
|||
|
|
<td>V/V</td>
|
|||
|
|
<td>AVX512VL AVX512BW</td>
|
|||
|
|
<td>Converts 16 packed word integers from ymm2 into 16 packed bytes in xmm1/m128 with truncation under writemask k1.</td></tr>
|
|||
|
|
<tr>
|
|||
|
|
<td>EVEX.256.F3.0F38.W0 20 /r VPMOVSWB xmm1/m128 {k1}{z}, ymm2</td>
|
|||
|
|
<td>A</td>
|
|||
|
|
<td>V/V</td>
|
|||
|
|
<td>AVX512VL AVX512BW</td>
|
|||
|
|
<td>Converts 16 packed signed word integers from ymm2 into 16 packed signed bytes in xmm1/m128 using signed saturation under writemask k1.</td></tr>
|
|||
|
|
<tr>
|
|||
|
|
<td>EVEX.256.F3.0F38.W0 10 /r VPMOVUSWB xmm1/m128 {k1}{z}, ymm2</td>
|
|||
|
|
<td>A</td>
|
|||
|
|
<td>V/V</td>
|
|||
|
|
<td>AVX512VL AVX512BW</td>
|
|||
|
|
<td>Converts 16 packed unsigned word integers from ymm2 into 16 packed unsigned bytes in xmm1/m128 using unsigned saturation under writemask k1.</td></tr>
|
|||
|
|
<tr>
|
|||
|
|
<td>EVEX.512.F3.0F38.W0 30 /r VPMOVWB ymm1/m256 {k1}{z}, zmm2</td>
|
|||
|
|
<td>A</td>
|
|||
|
|
<td>V/V</td>
|
|||
|
|
<td>AVX512BW</td>
|
|||
|
|
<td>Converts 32 packed word integers from zmm2 into 32 packed bytes in ymm1/m256 with truncation under writemask k1.</td></tr>
|
|||
|
|
<tr>
|
|||
|
|
<td>EVEX.512.F3.0F38.W0 20 /r VPMOVSWB ymm1/m256 {k1}{z}, zmm2</td>
|
|||
|
|
<td>A</td>
|
|||
|
|
<td>V/V</td>
|
|||
|
|
<td>AVX512BW</td>
|
|||
|
|
<td>Converts 32 packed signed word integers from zmm2 into 32 packed signed bytes in ymm1/m256 using signed saturation under writemask k1.</td></tr>
|
|||
|
|
<tr>
|
|||
|
|
<td>EVEX.512.F3.0F38.W0 10 /r VPMOVUSWB ymm1/m256 {k1}{z}, zmm2</td>
|
|||
|
|
<td>A</td>
|
|||
|
|
<td>V/V</td>
|
|||
|
|
<td>AVX512BW</td>
|
|||
|
|
<td>Converts 32 packed unsigned word integers from zmm2 into 32 packed unsigned bytes in ymm1/m256 using unsigned saturation 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>Half Mem</td>
|
|||
|
|
<td>ModRM:r/m (w)</td>
|
|||
|
|
<td>ModRM:reg (r)</td>
|
|||
|
|
<td>N/A</td>
|
|||
|
|
<td>N/A</td></tr></table>
|
|||
|
|
<h3 id="description">Description<a class="anchor" href="#description">
|
|||
|
|
¶
|
|||
|
|
</a></h3>
|
|||
|
|
<p>VPMOVWB down converts 16-bit integers into packed bytes using truncation. VPMOVSWB converts signed 16-bit integers into packed signed bytes using signed saturation. VPMOVUSWB convert unsigned word values into unsigned byte values using unsigned saturation.</p>
|
|||
|
|
<p>The source operand is a ZMM/YMM/XMM register. The destination operand is a YMM/XMM/XMM register or a 256/128/64-bit memory location.</p>
|
|||
|
|
<p>Down-converted byte elements are written to the destination operand (the first operand) from the least-significant byte. Byte elements of the destination operand are updated according to the writemask. Bits (MAXVL-1:256/128/64) of the register destination are zeroed.</p>
|
|||
|
|
<p>Note: EVEX.vvvv is reserved and must be 1111b otherwise instructions will #UD.</p>
|
|||
|
|
<h3 id="operation">Operation<a class="anchor" href="#operation">
|
|||
|
|
¶
|
|||
|
|
</a></h3>
|
|||
|
|
<h4 id="vpmovwb-instruction--evex-encoded-versions--when-dest-is-a-register">VPMOVWB instruction (EVEX encoded versions) when dest is a register<a class="anchor" href="#vpmovwb-instruction--evex-encoded-versions--when-dest-is-a-register">
|
|||
|
|
¶
|
|||
|
|
</a></h4>
|
|||
|
|
<pre>(KL, VL) = (8, 128), (16, 256), (32, 512)
|
|||
|
|
FOR j := 0 TO Kl-1
|
|||
|
|
i := j * 8
|
|||
|
|
m := j * 16
|
|||
|
|
IF k1[j] OR *no writemask*
|
|||
|
|
THEN DEST[i+7:i] := TruncateWordToByte (SRC[m+15:m])
|
|||
|
|
ELSE
|
|||
|
|
IF *merging-masking* ; merging-masking
|
|||
|
|
THEN *DEST[i+7:i] remains unchanged*
|
|||
|
|
ELSE *zeroing-masking*
|
|||
|
|
; zeroing-masking
|
|||
|
|
DEST[i+7:i] = 0
|
|||
|
|
FI
|
|||
|
|
FI;
|
|||
|
|
ENDFOR
|
|||
|
|
DEST[MAXVL-1:VL/2] := 0;
|
|||
|
|
</pre>
|
|||
|
|
<h4 id="vpmovwb-instruction--evex-encoded-versions--when-dest-is-memory">VPMOVWB instruction (EVEX encoded versions) when dest is memory<a class="anchor" href="#vpmovwb-instruction--evex-encoded-versions--when-dest-is-memory">
|
|||
|
|
¶
|
|||
|
|
</a></h4>
|
|||
|
|
<pre>(KL, VL) = (8, 128), (16, 256), (32, 512)
|
|||
|
|
FOR j := 0 TO Kl-1
|
|||
|
|
i := j * 8
|
|||
|
|
m := j * 16
|
|||
|
|
IF k1[j] OR *no writemask*
|
|||
|
|
THEN DEST[i+7:i] := TruncateWordToByte (SRC[m+15:m])
|
|||
|
|
ELSE
|
|||
|
|
*DEST[i+7:i] remains unchanged* ; merging-masking
|
|||
|
|
FI;
|
|||
|
|
ENDFOR
|
|||
|
|
</pre>
|
|||
|
|
<h4 id="vpmovswb-instruction--evex-encoded-versions--when-dest-is-a-register">VPMOVSWB instruction (EVEX encoded versions) when dest is a register<a class="anchor" href="#vpmovswb-instruction--evex-encoded-versions--when-dest-is-a-register">
|
|||
|
|
¶
|
|||
|
|
</a></h4>
|
|||
|
|
<pre>(KL, VL) = (8, 128), (16, 256), (32, 512)
|
|||
|
|
FOR j := 0 TO Kl-1
|
|||
|
|
i := j * 8
|
|||
|
|
m := j * 16
|
|||
|
|
IF k1[j] OR *no writemask*
|
|||
|
|
THEN DEST[i+7:i] := SaturateSignedWordToByte (SRC[m+15:m])
|
|||
|
|
ELSE
|
|||
|
|
IF *merging-masking* ; merging-masking
|
|||
|
|
THEN *DEST[i+7:i] remains unchanged*
|
|||
|
|
ELSE *zeroing-masking*
|
|||
|
|
; zeroing-masking
|
|||
|
|
DEST[i+7:i] = 0
|
|||
|
|
FI
|
|||
|
|
FI;
|
|||
|
|
ENDFOR
|
|||
|
|
DEST[MAXVL-1:VL/2] := 0;
|
|||
|
|
</pre>
|
|||
|
|
<h4 id="vpmovswb-instruction--evex-encoded-versions--when-dest-is-memory">VPMOVSWB instruction (EVEX encoded versions) when dest is memory<a class="anchor" href="#vpmovswb-instruction--evex-encoded-versions--when-dest-is-memory">
|
|||
|
|
¶
|
|||
|
|
</a></h4>
|
|||
|
|
<pre>(KL, VL) = (8, 128), (16, 256), (32, 512)
|
|||
|
|
FOR j := 0 TO Kl-1
|
|||
|
|
i := j * 8
|
|||
|
|
m := j * 16
|
|||
|
|
IF k1[j] OR *no writemask*
|
|||
|
|
THEN DEST[i+7:i] := SaturateSignedWordToByte (SRC[m+15:m])
|
|||
|
|
ELSE
|
|||
|
|
*DEST[i+7:i] remains unchanged* ; merging-masking
|
|||
|
|
FI;
|
|||
|
|
ENDFOR
|
|||
|
|
</pre>
|
|||
|
|
<h4 id="vpmovuswb-instruction--evex-encoded-versions--when-dest-is-a-register">VPMOVUSWB instruction (EVEX encoded versions) when dest is a register<a class="anchor" href="#vpmovuswb-instruction--evex-encoded-versions--when-dest-is-a-register">
|
|||
|
|
¶
|
|||
|
|
</a></h4>
|
|||
|
|
<pre>(KL, VL) = (8, 128), (16, 256), (32, 512)
|
|||
|
|
FOR j := 0 TO Kl-1
|
|||
|
|
i := j * 8
|
|||
|
|
m := j * 16
|
|||
|
|
IF k1[j] OR *no writemask*
|
|||
|
|
THEN DEST[i+7:i] := SaturateUnsignedWordToByte (SRC[m+15:m])
|
|||
|
|
ELSE
|
|||
|
|
IF *merging-masking* ; merging-masking
|
|||
|
|
THEN *DEST[i+7:i] remains unchanged*
|
|||
|
|
ELSE *zeroing-masking*
|
|||
|
|
; zeroing-masking
|
|||
|
|
DEST[i+7:i] = 0
|
|||
|
|
FI
|
|||
|
|
FI;
|
|||
|
|
ENDFOR
|
|||
|
|
DEST[MAXVL-1:VL/2] := 0;
|
|||
|
|
</pre>
|
|||
|
|
<h4 id="vpmovuswb-instruction--evex-encoded-versions--when-dest-is-memory">VPMOVUSWB instruction (EVEX encoded versions) when dest is memory<a class="anchor" href="#vpmovuswb-instruction--evex-encoded-versions--when-dest-is-memory">
|
|||
|
|
¶
|
|||
|
|
</a></h4>
|
|||
|
|
<pre>(KL, VL) = (8, 128), (16, 256), (32, 512)
|
|||
|
|
FOR j := 0 TO Kl-1
|
|||
|
|
i := j * 8
|
|||
|
|
m := j * 16
|
|||
|
|
IF k1[j] OR *no writemask*
|
|||
|
|
THEN DEST[i+7:i] := SaturateUnsignedWordToByte (SRC[m+15:m])
|
|||
|
|
ELSE
|
|||
|
|
*DEST[i+7:i] remains unchanged* ; merging-masking
|
|||
|
|
FI;
|
|||
|
|
ENDFOR
|
|||
|
|
</pre>
|
|||
|
|
<h3 id="intel-c-c++-compiler-intrinsic-equivalents">Intel C/C++ Compiler Intrinsic Equivalents<a class="anchor" href="#intel-c-c++-compiler-intrinsic-equivalents">
|
|||
|
|
¶
|
|||
|
|
</a></h3>
|
|||
|
|
<pre>VPMOVUSWB __m256i _mm512_cvtusepi16_epi8(__m512i a);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB __m256i _mm512_mask_cvtusepi16_epi8(__m256i a, __mmask32 k, __m512i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB __m256i _mm512_maskz_cvtusepi16_epi8( __mmask32 k, __m512i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB void _mm512_mask_cvtusepi16_storeu_epi8(void * , __mmask32 k, __m512i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB __m256i _mm512_cvtsepi16_epi8(__m512i a);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB __m256i _mm512_mask_cvtsepi16_epi8(__m256i a, __mmask32 k, __m512i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB __m256i _mm512_maskz_cvtsepi16_epi8( __mmask32 k, __m512i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB void _mm512_mask_cvtsepi16_storeu_epi8(void * , __mmask32 k, __m512i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB __m256i _mm512_cvtepi16_epi8(__m512i a);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB __m256i _mm512_mask_cvtepi16_epi8(__m256i a, __mmask32 k, __m512i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB __m256i _mm512_maskz_cvtepi16_epi8( __mmask32 k, __m512i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB void _mm512_mask_cvtepi16_storeu_epi8(void * , __mmask32 k, __m512i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB __m128i _mm256_cvtusepi16_epi8(__m256i a);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB __m128i _mm256_mask_cvtusepi16_epi8(__m128i a, __mmask16 k, __m256i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB __m128i _mm256_maskz_cvtusepi16_epi8( __mmask16 k, __m256i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB void _mm256_mask_cvtusepi16_storeu_epi8(void * , __mmask16 k, __m256i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB __m128i _mm_cvtusepi16_epi8(__m128i a);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB __m128i _mm_mask_cvtusepi16_epi8(__m128i a, __mmask8 k, __m128i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB __m128i _mm_maskz_cvtusepi16_epi8( __mmask8 k, __m128i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVUSWB void _mm_mask_cvtusepi16_storeu_epi8(void * , __mmask8 k, __m128i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB __m128i _mm256_cvtsepi16_epi8(__m256i a);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB __m128i _mm256_mask_cvtsepi16_epi8(__m128i a, __mmask16 k, __m256i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB __m128i _mm256_maskz_cvtsepi16_epi8( __mmask16 k, __m256i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB void _mm256_mask_cvtsepi16_storeu_epi8(void * , __mmask16 k, __m256i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB __m128i _mm_cvtsepi16_epi8(__m128i a);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB __m128i _mm_mask_cvtsepi16_epi8(__m128i a, __mmask8 k, __m128i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB __m128i _mm_maskz_cvtsepi16_epi8( __mmask8 k, __m128i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVSWB void _mm_mask_cvtsepi16_storeu_epi8(void * , __mmask8 k, __m128i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB __m128i _mm256_cvtepi16_epi8(__m256i a);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB __m128i _mm256_mask_cvtepi16_epi8(__m128i a, __mmask16 k, __m256i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB __m128i _mm256_maskz_cvtepi16_epi8( __mmask16 k, __m256i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB void _mm256_mask_cvtepi16_storeu_epi8(void * , __mmask16 k, __m256i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB __m128i _mm_cvtepi16_epi8(__m128i a);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB __m128i _mm_mask_cvtepi16_epi8(__m128i a, __mmask8 k, __m128i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB __m128i _mm_maskz_cvtepi16_epi8( __mmask8 k, __m128i b);
|
|||
|
|
</pre>
|
|||
|
|
<pre>VPMOVWB void _mm_mask_cvtepi16_storeu_epi8(void * , __mmask8 k, __m128i b);
|
|||
|
|
</pre>
|
|||
|
|
<h3 class="exceptions" id="simd-floating-point-exceptions">SIMD Floating-Point Exceptions<a class="anchor" href="#simd-floating-point-exceptions">
|
|||
|
|
¶
|
|||
|
|
</a></h3>
|
|||
|
|
<p>None.</p>
|
|||
|
|
<h3 class="exceptions" id="other-exceptions">Other Exceptions<a class="anchor" href="#other-exceptions">
|
|||
|
|
¶
|
|||
|
|
</a></h3>
|
|||
|
|
<p>EVEX-encoded instruction, see <span class="not-imported">Table 2-53</span>, “Type E6 Class Exception Conditions.”</p>
|
|||
|
|
<p>Additionally:</p>
|
|||
|
|
<table>
|
|||
|
|
<tr>
|
|||
|
|
<td>#UD</td>
|
|||
|
|
<td>If EVEX.vvvv != 1111B.</td></tr></table><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>
|