Lines Matching +full:0 +full:- +full:128
1 /*===---- smmintrin.h - SSE4 intrinsics ------------------------------------===
3 * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
5 * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
7 *===-----------------------------------------------------------------------===
22 __target__("sse4.1,no-evex512"), __min_vector_width__(128)))
25 #define _MM_FROUND_TO_NEAREST_INT 0x00
26 #define _MM_FROUND_TO_NEG_INF 0x01
27 #define _MM_FROUND_TO_POS_INF 0x02
28 #define _MM_FROUND_TO_ZERO 0x03
29 #define _MM_FROUND_CUR_DIRECTION 0x04
31 #define _MM_FROUND_RAISE_EXC 0x00
32 #define _MM_FROUND_NO_EXC 0x08
41 /// Rounds up each element of the 128-bit vector of [4 x float] to an
42 /// integer and returns the rounded values in a 128-bit vector of
54 /// A 128-bit vector of [4 x float] values to be rounded up.
55 /// \returns A 128-bit vector of [4 x float] containing the rounded values.
58 /// Rounds up each element of the 128-bit vector of [2 x double] to an
59 /// integer and returns the rounded values in a 128-bit vector of
71 /// A 128-bit vector of [2 x double] values to be rounded up.
72 /// \returns A 128-bit vector of [2 x double] containing the rounded values.
75 /// Copies three upper elements of the first 128-bit vector operand to
76 /// the corresponding three upper elements of the 128-bit result vector of
77 /// [4 x float]. Rounds up the lowest element of the second 128-bit vector
78 /// operand to an integer and copies it to the lowest element of the 128-bit
90 /// A 128-bit vector of [4 x float]. The values stored in bits [127:32] are
93 /// A 128-bit vector of [4 x float]. The value stored in bits [31:0] is
96 /// \returns A 128-bit vector of [4 x float] containing the copied and rounded
100 /// Copies the upper element of the first 128-bit vector operand to the
101 /// corresponding upper element of the 128-bit result vector of [2 x double].
102 /// Rounds up the lower element of the second 128-bit vector operand to an
103 /// integer and copies it to the lower element of the 128-bit result vector
115 /// A 128-bit vector of [2 x double]. The value stored in bits [127:64] is
118 /// A 128-bit vector of [2 x double]. The value stored in bits [63:0] is
121 /// \returns A 128-bit vector of [2 x double] containing the copied and rounded
125 /// Rounds down each element of the 128-bit vector of [4 x float] to an
126 /// an integer and returns the rounded values in a 128-bit vector of
138 /// A 128-bit vector of [4 x float] values to be rounded down.
139 /// \returns A 128-bit vector of [4 x float] containing the rounded values.
142 /// Rounds down each element of the 128-bit vector of [2 x double] to an
143 /// integer and returns the rounded values in a 128-bit vector of
155 /// A 128-bit vector of [2 x double].
156 /// \returns A 128-bit vector of [2 x double] containing the rounded values.
159 /// Copies three upper elements of the first 128-bit vector operand to
160 /// the corresponding three upper elements of the 128-bit result vector of
161 /// [4 x float]. Rounds down the lowest element of the second 128-bit vector
162 /// operand to an integer and copies it to the lowest element of the 128-bit
174 /// A 128-bit vector of [4 x float]. The values stored in bits [127:32] are
177 /// A 128-bit vector of [4 x float]. The value stored in bits [31:0] is
180 /// \returns A 128-bit vector of [4 x float] containing the copied and rounded
184 /// Copies the upper element of the first 128-bit vector operand to the
185 /// corresponding upper element of the 128-bit result vector of [2 x double].
186 /// Rounds down the lower element of the second 128-bit vector operand to an
187 /// integer and copies it to the lower element of the 128-bit result vector
199 /// A 128-bit vector of [2 x double]. The value stored in bits [127:64] is
202 /// A 128-bit vector of [2 x double]. The value stored in bits [63:0] is
205 /// \returns A 128-bit vector of [2 x double] containing the copied and rounded
209 /// Rounds each element of the 128-bit vector of [4 x float] to an
211 /// argument and returns the rounded values in a 128-bit vector of
223 /// A 128-bit vector of [4 x float].
228 /// 0: A normal PE exception is used \n
231 /// 0: Use bits [1:0] of \a M \n
233 /// Bits [1:0] contain the rounding control definition: \n
238 /// \returns A 128-bit vector of [4 x float] containing the rounded values.
242 /// Copies three upper elements of the first 128-bit vector operand to
243 /// the corresponding three upper elements of the 128-bit result vector of
244 /// [4 x float]. Rounds the lowest element of the second 128-bit vector
246 /// by the third argument and copies it to the lowest element of the 128-bit
258 /// A 128-bit vector of [4 x float]. The values stored in bits [127:32] are
261 /// A 128-bit vector of [4 x float]. The value stored in bits [31:0] is
268 /// 0: A normal PE exception is used \n
271 /// 0: Use bits [1:0] of \a M \n
273 /// Bits [1:0] contain the rounding control definition: \n
278 /// \returns A 128-bit vector of [4 x float] containing the copied and rounded
284 /// Rounds each element of the 128-bit vector of [2 x double] to an
286 /// argument and returns the rounded values in a 128-bit vector of
298 /// A 128-bit vector of [2 x double].
303 /// 0: A normal PE exception is used \n
306 /// 0: Use bits [1:0] of \a M \n
308 /// Bits [1:0] contain the rounding control definition: \n
313 /// \returns A 128-bit vector of [2 x double] containing the rounded values.
317 /// Copies the upper element of the first 128-bit vector operand to the
318 /// corresponding upper element of the 128-bit result vector of [2 x double].
319 /// Rounds the lower element of the second 128-bit vector operand to an
321 /// argument and copies it to the lower element of the 128-bit result vector
333 /// A 128-bit vector of [2 x double]. The value stored in bits [127:64] is
336 /// A 128-bit vector of [2 x double]. The value stored in bits [63:0] is
343 /// 0: A normal PE exception is used \n
346 /// 0: Use bits [1:0] of \a M \n
348 /// Bits [1:0] contain the rounding control definition: \n
353 /// \returns A 128-bit vector of [2 x double] containing the copied and rounded
360 /// Returns a 128-bit vector of [2 x double] where the values are
373 /// A 128-bit vector of [2 x double].
375 /// A 128-bit vector of [2 x double].
377 /// An immediate integer operand, with mask bits [1:0] specifying how the
379 /// index of a copied value. When a mask bit is 0, the corresponding 64-bit
381 /// When a mask bit is 1, the corresponding 64-bit element in operand \a V2
383 /// \returns A 128-bit vector of [2 x double] containing the copied values.
388 /// Returns a 128-bit vector of [4 x float] where the values are selected
401 /// A 128-bit vector of [4 x float].
403 /// A 128-bit vector of [4 x float].
405 /// An immediate integer operand, with mask bits [3:0] specifying how the
407 /// index of a copied value. When a mask bit is 0, the corresponding 32-bit
409 /// When a mask bit is 1, the corresponding 32-bit element in operand \a V2
411 /// \returns A 128-bit vector of [4 x float] containing the copied values.
416 /// Returns a 128-bit vector of [2 x double] where the values are
425 /// A 128-bit vector of [2 x double].
427 /// A 128-bit vector of [2 x double].
429 /// A 128-bit vector operand, with mask bits 127 and 63 specifying how the
431 /// most significant bit of a copied value. When a mask bit is 0, the
432 /// corresponding 64-bit element in operand \a __V1 is copied to the same
433 /// position in the result. When a mask bit is 1, the corresponding 64-bit
435 /// \returns A 128-bit vector of [2 x double] containing the copied values.
443 /// Returns a 128-bit vector of [4 x float] where the values are
452 /// A 128-bit vector of [4 x float].
454 /// A 128-bit vector of [4 x float].
456 /// A 128-bit vector operand, with mask bits 127, 95, 63, and 31 specifying
458 /// to the most significant bit of a copied value. When a mask bit is 0, the
459 /// corresponding 32-bit element in operand \a __V1 is copied to the same
460 /// position in the result. When a mask bit is 1, the corresponding 32-bit
462 /// \returns A 128-bit vector of [4 x float] containing the copied values.
470 /// Returns a 128-bit vector of [16 x i8] where the values are selected
479 /// A 128-bit vector of [16 x i8].
481 /// A 128-bit vector of [16 x i8].
483 /// A 128-bit vector operand, with mask bits 127, 119, 111...7 specifying
485 /// to the most significant bit of a copied value. When a mask bit is 0, the
486 /// corresponding 8-bit element in operand \a __V1 is copied to the same
487 /// position in the result. When a mask bit is 1, the corresponding 8-bit
489 /// \returns A 128-bit vector of [16 x i8] containing the copied values.
497 /// Returns a 128-bit vector of [8 x i16] where the values are selected
510 /// A 128-bit vector of [8 x i16].
512 /// A 128-bit vector of [8 x i16].
514 /// An immediate integer operand, with mask bits [7:0] specifying how the
516 /// index of a copied value. When a mask bit is 0, the corresponding 16-bit
518 /// When a mask bit is 1, the corresponding 16-bit element in operand \a V2
520 /// \returns A 128-bit vector of [8 x i16] containing the copied values.
526 /// Multiples corresponding elements of two 128-bit vectors of [4 x i32]
527 /// and returns the lower 32 bits of the each product in a 128-bit vector of
535 /// A 128-bit integer vector.
537 /// A 128-bit integer vector.
538 /// \returns A 128-bit integer vector containing the products of both operands.
544 /// Multiplies corresponding even-indexed elements of two 128-bit
545 /// vectors of [4 x i32] and returns a 128-bit vector of [2 x i64]
553 /// A 128-bit vector of [4 x i32].
555 /// A 128-bit vector of [4 x i32].
556 /// \returns A 128-bit vector of [2 x i64] containing the products of both
564 /// Computes the dot product of the two 128-bit vectors of [4 x float]
565 /// and returns it in the elements of the 128-bit result vector of
581 /// A 128-bit vector of [4 x float].
583 /// A 128-bit vector of [4 x float].
590 /// is treated as zero. Bits [3:0] determine which elements of the result
591 /// will receive a copy of the final dot product, with bit [0] corresponding
595 /// \returns A 128-bit vector of [4 x float] containing the dot product.
599 /// Computes the dot product of the two 128-bit vectors of [2 x double]
600 /// and returns it in the elements of the 128-bit result vector of
616 /// A 128-bit vector of [2 x double].
618 /// A 128-bit vector of [2 x double].
625 /// is treated as zero. Bits [1:0] determine which elements of the result
626 /// will receive a copy of the final dot product, with bit [0] corresponding
635 /// Loads integer values from a 128-bit aligned memory location to a
636 /// 128-bit integer vector.
643 /// A pointer to a 128-bit aligned memory location that contains the integer
645 /// \returns A 128-bit integer vector containing the data stored at the
653 /// Compares the corresponding elements of two 128-bit vectors of
654 /// [16 x i8] and returns a 128-bit vector of [16 x i8] containing the lesser
662 /// A 128-bit vector of [16 x i8].
664 /// A 128-bit vector of [16 x i8]
665 /// \returns A 128-bit vector of [16 x i8] containing the lesser values.
671 /// Compares the corresponding elements of two 128-bit vectors of
672 /// [16 x i8] and returns a 128-bit vector of [16 x i8] containing the
680 /// A 128-bit vector of [16 x i8].
682 /// A 128-bit vector of [16 x i8].
683 /// \returns A 128-bit vector of [16 x i8] containing the greater values.
689 /// Compares the corresponding elements of two 128-bit vectors of
690 /// [8 x u16] and returns a 128-bit vector of [8 x u16] containing the lesser
698 /// A 128-bit vector of [8 x u16].
700 /// A 128-bit vector of [8 x u16].
701 /// \returns A 128-bit vector of [8 x u16] containing the lesser values.
707 /// Compares the corresponding elements of two 128-bit vectors of
708 /// [8 x u16] and returns a 128-bit vector of [8 x u16] containing the
716 /// A 128-bit vector of [8 x u16].
718 /// A 128-bit vector of [8 x u16].
719 /// \returns A 128-bit vector of [8 x u16] containing the greater values.
725 /// Compares the corresponding elements of two 128-bit vectors of
726 /// [4 x i32] and returns a 128-bit vector of [4 x i32] containing the lesser
734 /// A 128-bit vector of [4 x i32].
736 /// A 128-bit vector of [4 x i32].
737 /// \returns A 128-bit vector of [4 x i32] containing the lesser values.
743 /// Compares the corresponding elements of two 128-bit vectors of
744 /// [4 x i32] and returns a 128-bit vector of [4 x i32] containing the
752 /// A 128-bit vector of [4 x i32].
754 /// A 128-bit vector of [4 x i32].
755 /// \returns A 128-bit vector of [4 x i32] containing the greater values.
761 /// Compares the corresponding elements of two 128-bit vectors of
762 /// [4 x u32] and returns a 128-bit vector of [4 x u32] containing the lesser
770 /// A 128-bit vector of [4 x u32].
772 /// A 128-bit vector of [4 x u32].
773 /// \returns A 128-bit vector of [4 x u32] containing the lesser values.
779 /// Compares the corresponding elements of two 128-bit vectors of
780 /// [4 x u32] and returns a 128-bit vector of [4 x u32] containing the
788 /// A 128-bit vector of [4 x u32].
790 /// A 128-bit vector of [4 x u32].
791 /// \returns A 128-bit vector of [4 x u32] containing the greater values.
801 /// resulting 128-bit vector of [4 x float] is then returned.
812 /// A 128-bit vector source operand of [4 x float]. With the exception of
814 /// [3:0] of \a N, all bits from this parameter are copied to the result.
816 /// A 128-bit vector source operand of [4 x float]. One single-precision
817 /// floating-point element from this source, as determined by the immediate
824 /// 00: Selects bits [31:0] from operand \a Y. \n
830 /// 00: Copies the selected bits from \a Y to result bits [31:0]. \n
834 /// Bits[3:0]: If any of these bits are set, the corresponding result
836 /// \returns A 128-bit vector of [4 x float] containing the copied
837 /// single-precision floating point elements from the operands.
840 /// Extracts a 32-bit integer from a 128-bit vector of [4 x float] and
853 /// A 128-bit vector of [4 x float].
855 /// An immediate value. Bits [1:0] determines which bits from the argument
857 /// 00: Bits [31:0] of parameter \a X are returned. \n
861 /// \returns A 32-bit integer containing the extracted 32 bits of float data.
867 /* Extract a single-precision float from X at index N into D. */
871 } while (0)
879 _mm_insert_ps(_mm_setzero_ps(), (X), _MM_MK_INSERTPS_NDX((N), 0, 0x0e))
882 /// Constructs a 128-bit vector of [16 x i8] by first making a copy of
883 /// the 128-bit integer vector parameter, and then inserting the lower 8 bits
896 /// A 128-bit integer vector of [16 x i8]. This vector is copied to the
903 /// An immediate value. Bits [3:0] specify the bit offset in the result at
905 /// 0000: Bits [7:0] of the result are used for insertion. \n
921 /// \returns A 128-bit integer vector containing the constructed values.
926 /// Constructs a 128-bit vector of [4 x i32] by first making a copy of
927 /// the 128-bit integer vector parameter, and then inserting the 32-bit
940 /// A 128-bit integer vector of [4 x i32]. This vector is copied to the
944 /// A 32-bit integer that is written to the result beginning at the offset
947 /// An immediate value. Bits [1:0] specify the bit offset in the result at
949 /// 00: Bits [31:0] of the result are used for insertion. \n
953 /// \returns A 128-bit integer vector containing the constructed values.
959 /// Constructs a 128-bit vector of [2 x i64] by first making a copy of
960 /// the 128-bit integer vector parameter, and then inserting the 64-bit
973 /// A 128-bit integer vector of [2 x i64]. This vector is copied to the
977 /// A 64-bit integer that is written to the result beginning at the offset
980 /// An immediate value. Bit [0] specifies the bit offset in the result at
982 /// 0: Bits [63:0] of the result are used for insertion. \n
984 /// \returns A 128-bit integer vector containing the constructed values.
993 /// Extracts an 8-bit element from the 128-bit integer vector of
1005 /// A 128-bit integer vector.
1007 /// An immediate value. Bits [3:0] specify which 8-bit vector element from
1009 /// 0000: Bits [7:0] of parameter \a X are extracted. \n
1026 /// 128-bit integer vector parameter and the remaining bits are assigned
1032 /// Extracts a 32-bit element from the 128-bit integer vector of
1044 /// A 128-bit integer vector.
1046 /// An immediate value. Bits [1:0] specify which 32-bit vector element from
1048 /// 00: Bits [31:0] of the parameter \a X are extracted. \n
1052 /// \returns An integer, whose lower 32 bits are selected from the 128-bit
1057 /// Extracts a 64-bit element from the 128-bit integer vector of
1067 /// in 64-bit mode.
1070 /// A 128-bit integer vector.
1072 /// An immediate value. Bit [0] specifies which 64-bit vector element from
1074 /// 0: Bits [63:0] are returned. \n
1076 /// \returns A 64-bit integer.
1080 /* SSE4 128-bit Packed Integer Comparisons. */
1081 /// Tests whether the specified bits in a 128-bit integer vector are all
1089 /// A 128-bit integer vector containing the bits to be tested.
1091 /// A 128-bit integer vector selecting which bits to test in operand \a __M.
1098 /// Tests whether the specified bits in a 128-bit integer vector are all
1106 /// A 128-bit integer vector containing the bits to be tested.
1108 /// A 128-bit integer vector selecting which bits to test in operand \a __M.
1115 /// Tests whether the specified bits in a 128-bit integer vector are
1123 /// A 128-bit integer vector containing the bits to be tested.
1125 /// A 128-bit integer vector selecting which bits to test in operand \a __M.
1133 /// Tests whether the specified bits in a 128-bit integer vector are all
1145 /// A 128-bit integer vector containing the bits to be tested.
1148 #define _mm_test_all_ones(V) _mm_testc_si128((V), _mm_set1_epi32(-1))
1150 /// Tests whether the specified bits in a 128-bit integer vector are
1162 /// A 128-bit integer vector containing the bits to be tested.
1164 /// A 128-bit integer vector selecting which bits to test in operand \a M.
1169 /// Tests whether the specified bits in a 128-bit integer vector are all
1181 /// A 128-bit integer vector containing the bits to be tested.
1183 /// A 128-bit integer vector selecting which bits to test in operand \a M.
1187 /* SSE4 64-bit Packed Integer Comparisons. */
1188 /// Compares each of the corresponding 64-bit values of the 128-bit
1191 /// Each comparison returns 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
1198 /// A 128-bit integer vector.
1200 /// A 128-bit integer vector.
1201 /// \returns A 128-bit integer vector containing the comparison results.
1207 /* SSE4 Packed Integer Sign-Extension. */
1208 /// Sign-extends each of the lower eight 8-bit integer elements of a
1209 /// 128-bit vector of [16 x i8] to 16-bit values and returns them in a
1210 /// 128-bit vector of [8 x i16]. The upper eight elements of the input vector
1218 /// A 128-bit vector of [16 x i8]. The lower eight 8-bit elements are
1219 /// sign-extended to 16-bit values.
1220 /// \returns A 128-bit vector of [8 x i16] containing the sign-extended values.
1225 __builtin_shufflevector((__v16qs)__V, (__v16qs)__V, 0, 1, 2, 3, 4, 5, 6, in _mm_cvtepi8_epi16()
1230 /// Sign-extends each of the lower four 8-bit integer elements of a
1231 /// 128-bit vector of [16 x i8] to 32-bit values and returns them in a
1232 /// 128-bit vector of [4 x i32]. The upper twelve elements of the input
1240 /// A 128-bit vector of [16 x i8]. The lower four 8-bit elements are
1241 /// sign-extended to 32-bit values.
1242 /// \returns A 128-bit vector of [4 x i32] containing the sign-extended values.
1247 __builtin_shufflevector((__v16qs)__V, (__v16qs)__V, 0, 1, 2, 3), __v4si); in _mm_cvtepi8_epi32()
1250 /// Sign-extends each of the lower two 8-bit integer elements of a
1251 /// 128-bit integer vector of [16 x i8] to 64-bit values and returns them in
1252 /// a 128-bit vector of [2 x i64]. The upper fourteen elements of the input
1260 /// A 128-bit vector of [16 x i8]. The lower two 8-bit elements are
1261 /// sign-extended to 64-bit values.
1262 /// \returns A 128-bit vector of [2 x i64] containing the sign-extended values.
1267 __builtin_shufflevector((__v16qs)__V, (__v16qs)__V, 0, 1), __v2di); in _mm_cvtepi8_epi64()
1270 /// Sign-extends each of the lower four 16-bit integer elements of a
1271 /// 128-bit integer vector of [8 x i16] to 32-bit values and returns them in
1272 /// a 128-bit vector of [4 x i32]. The upper four elements of the input
1280 /// A 128-bit vector of [8 x i16]. The lower four 16-bit elements are
1281 /// sign-extended to 32-bit values.
1282 /// \returns A 128-bit vector of [4 x i32] containing the sign-extended values.
1285 __builtin_shufflevector((__v8hi)__V, (__v8hi)__V, 0, 1, 2, 3), __v4si); in _mm_cvtepi16_epi32()
1288 /// Sign-extends each of the lower two 16-bit integer elements of a
1289 /// 128-bit integer vector of [8 x i16] to 64-bit values and returns them in
1290 /// a 128-bit vector of [2 x i64]. The upper six elements of the input
1298 /// A 128-bit vector of [8 x i16]. The lower two 16-bit elements are
1299 /// sign-extended to 64-bit values.
1300 /// \returns A 128-bit vector of [2 x i64] containing the sign-extended values.
1303 __builtin_shufflevector((__v8hi)__V, (__v8hi)__V, 0, 1), __v2di); in _mm_cvtepi16_epi64()
1306 /// Sign-extends each of the lower two 32-bit integer elements of a
1307 /// 128-bit integer vector of [4 x i32] to 64-bit values and returns them in
1308 /// a 128-bit vector of [2 x i64]. The upper two elements of the input vector
1316 /// A 128-bit vector of [4 x i32]. The lower two 32-bit elements are
1317 /// sign-extended to 64-bit values.
1318 /// \returns A 128-bit vector of [2 x i64] containing the sign-extended values.
1321 __builtin_shufflevector((__v4si)__V, (__v4si)__V, 0, 1), __v2di); in _mm_cvtepi32_epi64()
1324 /* SSE4 Packed Integer Zero-Extension. */
1325 /// Zero-extends each of the lower eight 8-bit integer elements of a
1326 /// 128-bit vector of [16 x i8] to 16-bit values and returns them in a
1327 /// 128-bit vector of [8 x i16]. The upper eight elements of the input vector
1335 /// A 128-bit vector of [16 x i8]. The lower eight 8-bit elements are
1336 /// zero-extended to 16-bit values.
1337 /// \returns A 128-bit vector of [8 x i16] containing the zero-extended values.
1340 __builtin_shufflevector((__v16qu)__V, (__v16qu)__V, 0, 1, 2, 3, 4, 5, 6, in _mm_cvtepu8_epi16()
1345 /// Zero-extends each of the lower four 8-bit integer elements of a
1346 /// 128-bit vector of [16 x i8] to 32-bit values and returns them in a
1347 /// 128-bit vector of [4 x i32]. The upper twelve elements of the input
1355 /// A 128-bit vector of [16 x i8]. The lower four 8-bit elements are
1356 /// zero-extended to 32-bit values.
1357 /// \returns A 128-bit vector of [4 x i32] containing the zero-extended values.
1360 __builtin_shufflevector((__v16qu)__V, (__v16qu)__V, 0, 1, 2, 3), __v4si); in _mm_cvtepu8_epi32()
1363 /// Zero-extends each of the lower two 8-bit integer elements of a
1364 /// 128-bit integer vector of [16 x i8] to 64-bit values and returns them in
1365 /// a 128-bit vector of [2 x i64]. The upper fourteen elements of the input
1373 /// A 128-bit vector of [16 x i8]. The lower two 8-bit elements are
1374 /// zero-extended to 64-bit values.
1375 /// \returns A 128-bit vector of [2 x i64] containing the zero-extended values.
1378 __builtin_shufflevector((__v16qu)__V, (__v16qu)__V, 0, 1), __v2di); in _mm_cvtepu8_epi64()
1381 /// Zero-extends each of the lower four 16-bit integer elements of a
1382 /// 128-bit integer vector of [8 x i16] to 32-bit values and returns them in
1383 /// a 128-bit vector of [4 x i32]. The upper four elements of the input
1391 /// A 128-bit vector of [8 x i16]. The lower four 16-bit elements are
1392 /// zero-extended to 32-bit values.
1393 /// \returns A 128-bit vector of [4 x i32] containing the zero-extended values.
1396 __builtin_shufflevector((__v8hu)__V, (__v8hu)__V, 0, 1, 2, 3), __v4si); in _mm_cvtepu16_epi32()
1399 /// Zero-extends each of the lower two 16-bit integer elements of a
1400 /// 128-bit integer vector of [8 x i16] to 64-bit values and returns them in
1401 /// a 128-bit vector of [2 x i64]. The upper six elements of the input vector
1409 /// A 128-bit vector of [8 x i16]. The lower two 16-bit elements are
1410 /// zero-extended to 64-bit values.
1411 /// \returns A 128-bit vector of [2 x i64] containing the zero-extended values.
1414 __builtin_shufflevector((__v8hu)__V, (__v8hu)__V, 0, 1), __v2di); in _mm_cvtepu16_epi64()
1417 /// Zero-extends each of the lower two 32-bit integer elements of a
1418 /// 128-bit integer vector of [4 x i32] to 64-bit values and returns them in
1419 /// a 128-bit vector of [2 x i64]. The upper two elements of the input vector
1427 /// A 128-bit vector of [4 x i32]. The lower two 32-bit elements are
1428 /// zero-extended to 64-bit values.
1429 /// \returns A 128-bit vector of [2 x i64] containing the zero-extended values.
1432 __builtin_shufflevector((__v4su)__V, (__v4su)__V, 0, 1), __v2di); in _mm_cvtepu32_epi64()
1436 /// Converts, with saturation, 32-bit signed integers from both 128-bit integer
1437 /// vector operands into 16-bit unsigned integers, and returns the packed
1440 /// Values greater than 0xFFFF are saturated to 0xFFFF. Values less than
1441 /// 0x0000 are saturated to 0x0000.
1448 /// A 128-bit vector of [4 x i32]. The converted [4 x i16] values are
1451 /// A 128-bit vector of [4 x i32]. The converted [4 x i16] values are
1453 /// \returns A 128-bit vector of [8 x i16] containing the converted values.
1460 /// Subtracts 8-bit unsigned integer values and computes the absolute
1474 /// A 128-bit vector of [16 x i8].
1476 /// A 128-bit vector of [16 x i8].
1478 /// An 8-bit immediate operand specifying how the absolute differences are to
1482 /// // M10 represents bits [1:0] of the immediate operand
1485 /// for (k = 0; k < 8; k = k + 1) {
1486 /// d0 = abs(X[i + k + 0] - Y[j + 0]);
1487 /// d1 = abs(X[i + k + 1] - Y[j + 1]);
1488 /// d2 = abs(X[i + k + 2] - Y[j + 2]);
1489 /// d3 = abs(X[i + k + 3] - Y[j + 3]);
1493 /// \returns A 128-bit integer vector containing the sums of the sets of
1499 /// Finds the minimum unsigned 16-bit element in the input 128-bit
1508 /// A 128-bit vector of [8 x u16].
1509 /// \returns A 128-bit value where bits [15:0] contain the minimum value found
1511 /// and the remaining bits are set to 0.
1526 #define _SIDD_UBYTE_OPS 0x00
1527 #define _SIDD_UWORD_OPS 0x01
1528 #define _SIDD_SBYTE_OPS 0x02
1529 #define _SIDD_SWORD_OPS 0x03
1532 #define _SIDD_CMP_EQUAL_ANY 0x00
1533 #define _SIDD_CMP_RANGES 0x04
1534 #define _SIDD_CMP_EQUAL_EACH 0x08
1535 #define _SIDD_CMP_EQUAL_ORDERED 0x0c
1538 #define _SIDD_POSITIVE_POLARITY 0x00
1539 #define _SIDD_NEGATIVE_POLARITY 0x10
1540 #define _SIDD_MASKED_POSITIVE_POLARITY 0x20
1541 #define _SIDD_MASKED_NEGATIVE_POLARITY 0x30
1544 #define _SIDD_LEAST_SIGNIFICANT 0x00
1545 #define _SIDD_MOST_SIGNIFICANT 0x40
1548 #define _SIDD_BIT_MASK 0x00
1549 #define _SIDD_UNIT_MASK 0x40
1554 /// \a A and \a B. Returns a 128-bit integer vector representing the result
1567 /// A 128-bit integer vector containing one of the source operands to be
1570 /// A 128-bit integer vector containing one of the source operands to be
1573 /// An 8-bit immediate operand specifying whether the characters are bytes or
1576 /// Bits [1:0]: Determine source data format. \n
1585 /// basis is greater than or equal for even-indexed elements in \a A,
1586 /// and less than or equal for odd-indexed elements in \a A. \n
1597 /// Bit [6]: Determines whether the result is zero-extended or expanded to 16
1599 /// 0: The result is zero-extended to 16 bytes. \n
1602 /// \returns Returns a 128-bit integer vector representing the result mask of
1623 /// A 128-bit integer vector containing one of the source operands to be
1626 /// A 128-bit integer vector containing one of the source operands to be
1629 /// An 8-bit immediate operand specifying whether the characters are bytes or
1632 /// Bits [1:0]: Determine source data format. \n
1641 /// basis is greater than or equal for even-indexed elements in \a A,
1642 /// and less than or equal for odd-indexed elements in \a A. \n
1655 /// 0: The index of the least significant set bit. \n
1664 /// \a A and \a B. Returns a 128-bit integer vector representing the result
1677 /// A 128-bit integer vector containing one of the source operands to be
1682 /// A 128-bit integer vector containing one of the source operands to be
1687 /// An 8-bit immediate operand specifying whether the characters are bytes or
1690 /// Bits [1:0]: Determine source data format. \n
1699 /// basis is greater than or equal for even-indexed elements in \a A,
1700 /// and less than or equal for odd-indexed elements in \a A. \n
1711 /// Bit [6]: Determines whether the result is zero-extended or expanded to 16
1713 /// 0: The result is zero-extended to 16 bytes. \n
1716 /// \returns Returns a 128-bit integer vector representing the result mask of
1738 /// A 128-bit integer vector containing one of the source operands to be
1743 /// A 128-bit integer vector containing one of the source operands to be
1748 /// An 8-bit immediate operand specifying whether the characters are bytes or
1751 /// Bits [1:0]: Determine source data format. \n
1760 /// basis is greater than or equal for even-indexed elements in \a A,
1761 /// and less than or equal for odd-indexed elements in \a A. \n
1774 /// 0: The index of the least significant set bit. \n
1786 /// string in \a B is the maximum, otherwise, returns 0.
1798 /// A 128-bit integer vector containing one of the source operands to be
1801 /// A 128-bit integer vector containing one of the source operands to be
1804 /// An 8-bit immediate operand specifying whether the characters are bytes or
1806 /// Bits [1:0]: Determine source data format. \n
1815 /// basis is greater than or equal for even-indexed elements in \a A,
1816 /// and less than or equal for odd-indexed elements in \a A. \n
1828 /// \a B is the maximum; otherwise, returns 0.
1835 /// \a A and \a B. Returns 1 if the bit mask is non-zero, otherwise, returns
1836 /// 0.
1848 /// A 128-bit integer vector containing one of the source operands to be
1851 /// A 128-bit integer vector containing one of the source operands to be
1854 /// An 8-bit immediate operand specifying whether the characters are bytes or
1856 /// Bits [1:0]: Determine source data format. \n
1865 /// basis is greater than or equal for even-indexed elements in \a A,
1866 /// and less than or equal for odd-indexed elements in \a A. \n
1877 /// \returns Returns 1 if the bit mask is non-zero, otherwise, returns 0.
1884 /// \a A and \a B. Returns bit 0 of the resulting bit mask.
1896 /// A 128-bit integer vector containing one of the source operands to be
1899 /// A 128-bit integer vector containing one of the source operands to be
1902 /// An 8-bit immediate operand specifying whether the characters are bytes or
1904 /// Bits [1:0]: Determine source data format. \n
1913 /// basis is greater than or equal for even-indexed elements in \a A,
1914 /// and less than or equal for odd-indexed elements in \a A. \n
1925 /// \returns Returns bit 0 of the resulting bit mask.
1933 /// the maximum, otherwise, returns 0.
1945 /// A 128-bit integer vector containing one of the source operands to be
1948 /// A 128-bit integer vector containing one of the source operands to be
1951 /// An 8-bit immediate operand specifying whether the characters are bytes or
1953 /// Bits [1:0]: Determine source data format. \n
1962 /// basis is greater than or equal for even-indexed elements in \a A,
1963 /// and less than or equal for odd-indexed elements in \a A. \n
1975 /// maximum, otherwise, returns 0.
1983 /// the maximum, otherwise, returns 0.
1995 /// A 128-bit integer vector containing one of the source operands to be
1998 /// A 128-bit integer vector containing one of the source operands to be
2001 /// An 8-bit immediate operand specifying whether the characters are bytes or
2003 /// Bits [1:0]: Determine source data format. \n
2012 /// basis is greater than or equal for even-indexed elements in \a A,
2013 /// and less than or equal for odd-indexed elements in \a A. \n
2025 /// maximum, otherwise, returns 0.
2033 /// string in \a B is the maximum, otherwise, returns 0.
2045 /// A 128-bit integer vector containing one of the source operands to be
2050 /// A 128-bit integer vector containing one of the source operands to be
2055 /// An 8-bit immediate operand specifying whether the characters are bytes or
2057 /// Bits [1:0]: Determine source data format. \n
2066 /// basis is greater than or equal for even-indexed elements in \a A,
2067 /// and less than or equal for odd-indexed elements in \a A. \n
2079 /// \a B is the maximum, otherwise, returns 0.
2087 /// \a A and \a B. Returns 1 if the resulting mask is non-zero, otherwise,
2088 /// returns 0.
2100 /// A 128-bit integer vector containing one of the source operands to be
2105 /// A 128-bit integer vector containing one of the source operands to be
2110 /// An 8-bit immediate operand specifying whether the characters are bytes or
2112 /// Bits [1:0]: Determine source data format. \n
2121 /// basis is greater than or equal for even-indexed elements in \a A,
2122 /// and less than or equal for odd-indexed elements in \a A. \n
2133 /// \returns Returns 1 if the resulting mask is non-zero, otherwise, returns 0.
2141 /// \a A and \a B. Returns bit 0 of the resulting bit mask.
2153 /// A 128-bit integer vector containing one of the source operands to be
2158 /// A 128-bit integer vector containing one of the source operands to be
2163 /// An 8-bit immediate operand specifying whether the characters are bytes or
2165 /// Bits [1:0]: Determine source data format. \n
2174 /// basis is greater than or equal for even-indexed elements in \a A,
2175 /// and less than or equal for odd-indexed elements in \a A. \n
2186 /// \returns Returns bit 0 of the resulting bit mask.
2195 /// the maximum, otherwise, returns 0.
2207 /// A 128-bit integer vector containing one of the source operands to be
2212 /// A 128-bit integer vector containing one of the source operands to be
2217 /// An 8-bit immediate operand specifying whether the characters are bytes or
2219 /// Bits [1:0]: Determine source data format. \n
2228 /// basis is greater than or equal for even-indexed elements in \a A,
2229 /// and less than or equal for odd-indexed elements in \a A. \n
2241 /// maximum, otherwise, returns 0.
2250 /// the maximum, otherwise, returns 0.
2261 /// A 128-bit integer vector containing one of the source operands to be
2266 /// A 128-bit integer vector containing one of the source operands to be
2271 /// An 8-bit immediate operand specifying whether the characters are bytes or
2273 /// Bits [1:0]: Determine source data format. \n
2282 /// basis is greater than or equal for even-indexed elements in \a A,
2283 /// and less than or equal for odd-indexed elements in \a A. \n
2295 /// maximum, otherwise, returns 0.
2301 /* SSE4.2 Compare Packed Data -- Greater Than. */
2302 /// Compares each of the corresponding 64-bit values of the 128-bit
2306 /// Each comparison returns 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
2313 /// A 128-bit integer vector.
2315 /// A 128-bit integer vector.
2316 /// \returns A 128-bit integer vector containing the comparison results.