Skip to content

Commit

Permalink
Merge commit '845e09b598cb4023a525badc5d63e6aaa136797f'
Browse files Browse the repository at this point in the history
  • Loading branch information
@milot-mirdita
milot-mirdita committed Nov 1, 2022
2 parents fce06b1 + 845e09b commit 2b4fc41
Show file tree
Hide file tree
Showing 250 changed files with 61,173 additions and 8,670 deletions.
40 changes: 40 additions & 0 deletions lib/simde/simde/arm/neon.h
View file
Expand Up @@ -34,6 +34,7 @@
#include "neon/abdl.h"
#include "neon/abs.h"
#include "neon/add.h"
#include "neon/addhn.h"
#include "neon/addl.h"
#include "neon/addlv.h"
#include "neon/addl_high.h"
Expand All @@ -58,6 +59,10 @@
#include "neon/clt.h"
#include "neon/cltz.h"
#include "neon/clz.h"
#include "neon/cmla.h"
#include "neon/cmla_rot90.h"
#include "neon/cmla_rot180.h"
#include "neon/cmla_rot270.h"
#include "neon/cnt.h"
#include "neon/cvt.h"
#include "neon/combine.h"
Expand All @@ -68,14 +73,21 @@
#include "neon/dup_n.h"
#include "neon/eor.h"
#include "neon/ext.h"
#include "neon/fma.h"
#include "neon/fma_lane.h"
#include "neon/fma_n.h"
#include "neon/get_high.h"
#include "neon/get_lane.h"
#include "neon/get_low.h"
#include "neon/hadd.h"
#include "neon/hsub.h"
#include "neon/ld1.h"
#include "neon/ld1_dup.h"
#include "neon/ld1_lane.h"
#include "neon/ld2.h"
#include "neon/ld3.h"
#include "neon/ld4.h"
#include "neon/ld4_lane.h"
#include "neon/max.h"
#include "neon/maxnm.h"
#include "neon/maxv.h"
Expand All @@ -86,10 +98,15 @@
#include "neon/mla_n.h"
#include "neon/mlal.h"
#include "neon/mlal_high.h"
#include "neon/mlal_high_n.h"
#include "neon/mlal_lane.h"
#include "neon/mlal_n.h"
#include "neon/mls.h"
#include "neon/mls_n.h"
#include "neon/mlsl.h"
#include "neon/mlsl_high.h"
#include "neon/mlsl_high_n.h"
#include "neon/mlsl_lane.h"
#include "neon/mlsl_n.h"
#include "neon/movl.h"
#include "neon/movl_high.h"
Expand All @@ -100,6 +117,7 @@
#include "neon/mul_n.h"
#include "neon/mull.h"
#include "neon/mull_high.h"
#include "neon/mull_lane.h"
#include "neon/mull_n.h"
#include "neon/mvn.h"
#include "neon/neg.h"
Expand All @@ -113,18 +131,28 @@
#include "neon/qabs.h"
#include "neon/qadd.h"
#include "neon/qdmulh.h"
#include "neon/qdmulh_lane.h"
#include "neon/qdmulh_n.h"
#include "neon/qdmull.h"
#include "neon/qrdmulh.h"
#include "neon/qrdmulh_lane.h"
#include "neon/qrdmulh_n.h"
#include "neon/qrshrn_n.h"
#include "neon/qrshrun_n.h"
#include "neon/qmovn.h"
#include "neon/qmovun.h"
#include "neon/qmovn_high.h"
#include "neon/qneg.h"
#include "neon/qsub.h"
#include "neon/qshl.h"
#include "neon/qshlu_n.h"
#include "neon/qshrn_n.h"
#include "neon/qshrun_n.h"
#include "neon/qtbl.h"
#include "neon/qtbx.h"
#include "neon/rbit.h"
#include "neon/recpe.h"
#include "neon/recps.h"
#include "neon/reinterpret.h"
#include "neon/rev16.h"
#include "neon/rev32.h"
Expand All @@ -137,19 +165,31 @@
#include "neon/rndp.h"
#include "neon/rshl.h"
#include "neon/rshr_n.h"
#include "neon/rshrn_n.h"
#include "neon/rsqrte.h"
#include "neon/rsqrts.h"
#include "neon/rsra_n.h"
#include "neon/set_lane.h"
#include "neon/shl.h"
#include "neon/shl_n.h"
#include "neon/shll_n.h"
#include "neon/shr_n.h"
#include "neon/shrn_n.h"
#include "neon/sqadd.h"
#include "neon/sra_n.h"
#include "neon/sri_n.h"
#include "neon/st1.h"
#include "neon/st1_lane.h"
#include "neon/st2.h"
#include "neon/st2_lane.h"
#include "neon/st3.h"
#include "neon/st3_lane.h"
#include "neon/st4.h"
#include "neon/st4_lane.h"
#include "neon/sub.h"
#include "neon/subhn.h"
#include "neon/subl.h"
#include "neon/subl_high.h"
#include "neon/subw.h"
#include "neon/subw_high.h"
#include "neon/tbl.h"
Expand Down
167 changes: 136 additions & 31 deletions lib/simde/simde/arm/neon/abd.h
View file
Expand Up @@ -100,6 +100,23 @@ simde_int8x8_t
simde_vabd_s8(simde_int8x8_t a, simde_int8x8_t b) {
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
return vabd_s8(a, b);
#elif defined(SIMDE_X86_MMX_NATIVE)
simde_int8x8_private
r_,
a_ = simde_int8x8_to_private(a),
b_ = simde_int8x8_to_private(b);

const __m64 m = _mm_cmpgt_pi8(b_.m64, a_.m64);
r_.m64 =
_mm_xor_si64(
_mm_add_pi8(
_mm_sub_pi8(a_.m64, b_.m64),
m
),
m
);

return simde_int8x8_from_private(r_);
#else
return simde_vmovn_s16(simde_vabsq_s16(simde_vsubl_s8(a, b)));
#endif
Expand All @@ -114,6 +131,15 @@ simde_int16x4_t
simde_vabd_s16(simde_int16x4_t a, simde_int16x4_t b) {
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
return vabd_s16(a, b);
#elif defined(SIMDE_X86_MMX_NATIVE) && defined(SIMDE_X86_SSE_NATIVE)
simde_int16x4_private
r_,
a_ = simde_int16x4_to_private(a),
b_ = simde_int16x4_to_private(b);

r_.m64 = _mm_sub_pi16(_mm_max_pi16(a_.m64, b_.m64), _mm_min_pi16(a_.m64, b_.m64));

return simde_int16x4_from_private(r_);
#else
return simde_vmovn_s32(simde_vabsq_s32(simde_vsubl_s16(a, b)));
#endif
Expand Down Expand Up @@ -227,17 +253,37 @@ simde_int8x16_t
simde_vabdq_s8(simde_int8x16_t a, simde_int8x16_t b) {
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
return vabdq_s8(a, b);
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
return vec_sub(vec_max(a, b), vec_min(a, b));
#elif defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
return vec_max(a, b) - vec_min(a, b);
#else
simde_int8x16_private
r_,
a_ = simde_int8x16_to_private(a),
b_ = simde_int8x16_to_private(b);

SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int16_t tmp = HEDLEY_STATIC_CAST(int16_t, a_.values[i]) - HEDLEY_STATIC_CAST(int16_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(int8_t, tmp < 0 ? -tmp : tmp);
}
#if defined(SIMDE_X86_SSE4_1_NATIVE)
r_.m128i = _mm_sub_epi8(_mm_max_epi8(a_.m128i, b_.m128i), _mm_min_epi8(a_.m128i, b_.m128i));
#elif defined(SIMDE_X86_SSE2_NATIVE)
const __m128i m = _mm_cmpgt_epi8(b_.m128i, a_.m128i);
r_.m128i =
_mm_xor_si128(
_mm_add_epi8(
_mm_sub_epi8(a_.m128i, b_.m128i),
m
),
m
);
#elif defined(SIMDE_WASM_SIMD128_NATIVE)
r_.v128 = wasm_i8x16_sub(wasm_i8x16_max(a_.v128, b_.v128), wasm_i8x16_min(a_.v128, b_.v128));
#else
SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int16_t tmp = HEDLEY_STATIC_CAST(int16_t, a_.values[i]) - HEDLEY_STATIC_CAST(int16_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(int8_t, tmp < 0 ? -tmp : tmp);
}
#endif

return simde_int8x16_from_private(r_);
#endif
Expand All @@ -252,18 +298,31 @@ simde_int16x8_t
simde_vabdq_s16(simde_int16x8_t a, simde_int16x8_t b) {
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
return vabdq_s16(a, b);
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
return vec_sub(vec_max(a, b), vec_min(a, b));
#elif defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
return vec_max(a, b) - vec_min(a, b);
#else
simde_int16x8_private
r_,
a_ = simde_int16x8_to_private(a),
b_ = simde_int16x8_to_private(b);

SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int32_t tmp = HEDLEY_STATIC_CAST(int32_t, a_.values[i]) - HEDLEY_STATIC_CAST(int32_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(int16_t, tmp < 0 ? -tmp : tmp);
}

#if defined(SIMDE_X86_SSE2_NATIVE)
/* https://github.com/simd-everywhere/simde/issues/855#issuecomment-881658604 */
r_.m128i = _mm_sub_epi16(_mm_max_epi16(a_.m128i, b_.m128i), _mm_min_epi16(a_.m128i, b_.m128i));
#elif defined(SIMDE_WASM_SIMD128_NATIVE)
r_.v128 = wasm_i16x8_sub(wasm_i16x8_max(a_.v128, b_.v128), wasm_i16x8_min(a_.v128, b_.v128));
#else
SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
r_.values[i] =
(a_.values[i] < b_.values[i]) ?
(b_.values[i] - a_.values[i]) :
(a_.values[i] - b_.values[i]);
}

#endif
return simde_int16x8_from_private(r_);
#endif
}
Expand All @@ -277,17 +336,35 @@ simde_int32x4_t
simde_vabdq_s32(simde_int32x4_t a, simde_int32x4_t b) {
#if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
return vabdq_s32(a, b);
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
return vec_sub(vec_max(a, b), vec_min(a, b));
#elif defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
return vec_max(a, b) - vec_min(a, b);
#else
simde_int32x4_private
r_,
a_ = simde_int32x4_to_private(a),
b_ = simde_int32x4_to_private(b);

SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int64_t tmp = HEDLEY_STATIC_CAST(int64_t, a_.values[i]) - HEDLEY_STATIC_CAST(int64_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(int32_t, tmp < 0 ? -tmp : tmp);
}
#if defined(SIMDE_X86_SSE4_1_NATIVE)
r_.m128i = _mm_sub_epi32(_mm_max_epi32(a_.m128i, b_.m128i), _mm_min_epi32(a_.m128i, b_.m128i));
#elif defined(SIMDE_X86_SSE2_NATIVE)
const __m128i m = _mm_cmpgt_epi32(b_.m128i, a_.m128i);
r_.m128i =
_mm_xor_si128(
_mm_add_epi32(
_mm_sub_epi32(a_.m128i, b_.m128i),
m
),
m
);
#else
SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int64_t tmp = HEDLEY_STATIC_CAST(int64_t, a_.values[i]) - HEDLEY_STATIC_CAST(int64_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(int32_t, tmp < 0 ? -tmp : tmp);
}
#endif

return simde_int32x4_from_private(r_);
#endif
Expand All @@ -304,17 +381,27 @@ simde_vabdq_u8(simde_uint8x16_t a, simde_uint8x16_t b) {
return vabdq_u8(a, b);
#elif defined(SIMDE_POWER_ALTIVEC_P9_NATIVE)
return vec_absd(a, b);
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
return vec_sub(vec_max(a, b), vec_min(a, b));
#elif defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
return vec_max(a, b) - vec_min(a, b);
#else
simde_uint8x16_private
r_,
a_ = simde_uint8x16_to_private(a),
b_ = simde_uint8x16_to_private(b);

SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int16_t tmp = HEDLEY_STATIC_CAST(int16_t, a_.values[i]) - HEDLEY_STATIC_CAST(int16_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(uint8_t, tmp < 0 ? -tmp : tmp);
}
#if defined(SIMDE_X86_SSE2_NATIVE)
r_.m128i = _mm_sub_epi8(_mm_max_epu8(a_.m128i, b_.m128i), _mm_min_epu8(a_.m128i, b_.m128i));
#elif defined(SIMDE_WASM_SIMD128_NATIVE)
r_.v128 = wasm_i8x16_sub(wasm_u8x16_max(a_.v128, b_.v128), wasm_u8x16_min(a_.v128, b_.v128));
#else
SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int16_t tmp = HEDLEY_STATIC_CAST(int16_t, a_.values[i]) - HEDLEY_STATIC_CAST(int16_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(uint8_t, tmp < 0 ? -tmp : tmp);
}
#endif

return simde_uint8x16_from_private(r_);
#endif
Expand All @@ -331,17 +418,27 @@ simde_vabdq_u16(simde_uint16x8_t a, simde_uint16x8_t b) {
return vabdq_u16(a, b);
#elif defined(SIMDE_POWER_ALTIVEC_P9_NATIVE)
return vec_absd(a, b);
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
return vec_sub(vec_max(a, b), vec_min(a, b));
#elif defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
return vec_max(a, b) - vec_min(a, b);
#else
simde_uint16x8_private
r_,
a_ = simde_uint16x8_to_private(a),
b_ = simde_uint16x8_to_private(b);

SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int32_t tmp = HEDLEY_STATIC_CAST(int32_t, a_.values[i]) - HEDLEY_STATIC_CAST(int32_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(uint16_t, tmp < 0 ? -tmp : tmp);
}
#if defined(SIMDE_X86_SSE4_2_NATIVE)
r_.m128i = _mm_sub_epi16(_mm_max_epu16(a_.m128i, b_.m128i), _mm_min_epu16(a_.m128i, b_.m128i));
#elif defined(SIMDE_WASM_SIMD128_NATIVE)
r_.v128 = wasm_i16x8_sub(wasm_u16x8_max(a_.v128, b_.v128), wasm_u16x8_min(a_.v128, b_.v128));
#else
SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int32_t tmp = HEDLEY_STATIC_CAST(int32_t, a_.values[i]) - HEDLEY_STATIC_CAST(int32_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(uint16_t, tmp < 0 ? -tmp : tmp);
}
#endif

return simde_uint16x8_from_private(r_);
#endif
Expand All @@ -358,17 +455,25 @@ simde_vabdq_u32(simde_uint32x4_t a, simde_uint32x4_t b) {
return vabdq_u32(a, b);
#elif defined(SIMDE_POWER_ALTIVEC_P9_NATIVE)
return vec_absd(a, b);
#elif defined(SIMDE_POWER_ALTIVEC_P6_NATIVE)
return vec_sub(vec_max(a, b), vec_min(a, b));
#elif defined(SIMDE_ZARCH_ZVECTOR_13_NATIVE)
return vec_max(a, b) - vec_min(a, b);
#else
simde_uint32x4_private
r_,
a_ = simde_uint32x4_to_private(a),
b_ = simde_uint32x4_to_private(b);

SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int64_t tmp = HEDLEY_STATIC_CAST(int64_t, a_.values[i]) - HEDLEY_STATIC_CAST(int64_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(uint32_t, tmp < 0 ? -tmp : tmp);
}
#if defined(SIMDE_X86_SSE4_2_NATIVE)
r_.m128i = _mm_sub_epi32(_mm_max_epu32(a_.m128i, b_.m128i), _mm_min_epu32(a_.m128i, b_.m128i));
#else
SIMDE_VECTORIZE
for (size_t i = 0 ; i < (sizeof(r_.values) / sizeof(r_.values[0])) ; i++) {
int64_t tmp = HEDLEY_STATIC_CAST(int64_t, a_.values[i]) - HEDLEY_STATIC_CAST(int64_t, b_.values[i]);
r_.values[i] = HEDLEY_STATIC_CAST(uint32_t, tmp < 0 ? -tmp : tmp);
}
#endif

return simde_uint32x4_from_private(r_);
#endif
Expand Down

0 comments on commit 2b4fc41

Please sign in to comment.