| /external/gemmlowp/internal/ |
| fixedpoint_neon.h | 28 inline int32x4_t BitAnd(int32x4_t a, int32x4_t b) {
33 inline int32x4_t BitOr(int32x4_t a, int32x4_t b) {
38 inline int32x4_t BitXor(int32x4_t a, int32x4_t b) {
43 inline int32x4_t BitNot(int32x4_t a)
[all...] |
| unpack_neon.h | 28 int32x4_t RoundingMultiplyByConstantFraction(int32x4_t x) {
45 const int32x4_t remaining_product =
52 int32x4_t get_int32x4_t_and_inc(
54 const int32x4_t result = vld1q_s32(iterator->get());
60 int32x4_t get_int32x4_t_and_inc(
62 const int32x4_t result = vdupq_n_s32(**iterator);
107 int32x4_t raw_xx[4];
112 int32x4_t raw_x1[4];
114 const int32x4_t sum_x1 = vld1q_s32(sums_of_each_slice_ptr)
[all...] |
| output_neon.h | 27 typedef Fragment<int32x4_t, 4, 1, MapOrder::ColMajor> NEONFragmentInt32x4x1;
38 // 4 int32x4_t components. This partial specialization handles that for
39 // arbitrary output stages implementing a int32x4_t path. Only some output
80 const int32x4_t a = vaddq_s32(input, vdupq_n_s32(result_offset));
81 const int32x4_t b =
106 const int32x4_t result_mult_int =
108 const int32x4_t result_offset =
110 const int32x4_t a = vaddq_s32(input, result_offset);
111 const int32x4_t b =
136 const int32x4_t result_mult_int
[all...] |
| /ndk/tests/build/b9193874-neon/jni/ |
| b9193874-neon.c | 9 typedef __builtin_neon_si int32x4_t __attribute__ ((__vector_size__ (16))); typedef
28 __extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
31 return (int32x4_t)__builtin_neon_vaddlv4hi (__a, __b, 1);
48 __extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
51 return (int32x4_t)__builtin_neon_vsublv4hi (__a, __b, 1);
64 vrshrn_n_s32 (int32x4_t __a, const int __b)
73 __extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
76 return (int32x4_t)__builtin_neon_vshll_nv4hi (__a, __b, 1);
113 __extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
116 return (int32x4_t)__builtin_neon_vmovlv4hi (__a, 1)
[all...] |
| /external/clang/test/CodeGen/ |
| arm64_neon_high_half.c | 11 int32x4_t test_vaddw_high_s16(int32x4_t lhs, int16x8_t rhs) {
16 int64x2_t test_vaddw_high_s32(int64x2_t lhs, int32x4_t rhs) {
41 int32x4_t test_vsubw_high_s16(int32x4_t lhs, int16x8_t rhs) {
46 int64x2_t test_vsubw_high_s32(int64x2_t lhs, int32x4_t rhs) {
71 int32x4_t test_vabdl_high_s16(int16x8_t lhs, int16x8_t rhs) {
76 int64x2_t test_vabdl_high_s32(int32x4_t lhs, int32x4_t rhs) {
101 int32x4_t test_vabal_high_s16(int32x4_t accum, int16x8_t lhs, int16x8_t rhs)
[all...] |
| aarch64-neon-2velem.c | 26 int32x4_t test_vmlaq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v) {
44 int32x2_t test_vmla_laneq_s32(int32x2_t a, int32x2_t b, int32x4_t v) {
50 int32x4_t test_vmlaq_laneq_s32(int32x4_t a, int32x4_t b, int32x4_t v) {
74 int32x4_t test_vmlsq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v)
[all...] |
| arm-v8.1a-neon-intrinsics.c | 33 int32x4_t test_vqrdmlahq_s32(int32x4_t a, int32x4_t b, int32x4_t c) {
61 int32x4_t test_vqrdmlahq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t c) {
89 int32x4_t test_vqrdmlshq_s32(int32x4_t a, int32x4_t b, int32x4_t c)
[all...] |
| arm64-vrnd.c | 7 int32x4_t rnd3(float32x4_t a) { return vrndq_f32(a); }
15 int32x4_t rnd8(float32x4_t a) { return vrndnq_f32(a); }
24 int32x4_t rnd12(float32x4_t a) { return vrndmq_f32(a); }
33 int32x4_t rnd16(float32x4_t a) { return vrndpq_f32(a); }
40 int32x4_t rnd20(float32x4_t a) { return vrndaq_f32(a); }
47 int32x4_t rnd24(float32x4_t a) { return vrndxq_f32(a); }
|
| aarch64-v8.1a-neon-intrinsics.c | 16 int32x2_t test_vqrdmlah_laneq_s32(int32x2_t a, int32x2_t b, int32x4_t v) {
28 int32x4_t test_vqrdmlahq_laneq_s32(int32x4_t a, int32x4_t b, int32x4_t v) {
64 int32_t test_vqrdmlahs_laneq_s32(int32_t a, int32_t b, int32x4_t c) {
76 int32x2_t test_vqrdmlsh_laneq_s32(int32x2_t a, int32x2_t b, int32x4_t v) {
88 int32x4_t test_vqrdmlshq_laneq_s32(int32x4_t a, int32x4_t b, int32x4_t v)
[all...] |
| /external/webrtc/webrtc/modules/audio_coding/codecs/isac/fix/source/ |
| transform_neon.c | 35 int32x4_t factq = vdupq_n_s32(fact);
50 int32x4_t tmp0 = vmull_s16(vget_low_s16(tmpr), vget_low_s16(inre1));
51 int32x4_t tmp1 = vmull_s16(vget_low_s16(tmpr), vget_low_s16(inre2));
55 int32x4_t tmp2 = vmull_high_s16(tmpr, inre1);
56 int32x4_t tmp3 = vmull_high_s16(tmpr, inre2);
60 int32x4_t tmp2 = vmull_s16(vget_high_s16(tmpr), vget_high_s16(inre1));
61 int32x4_t tmp3 = vmull_s16(vget_high_s16(tmpr), vget_high_s16(inre2));
66 int32x4_t outr_0 = vqdmulhq_s32(tmp0, factq);
67 int32x4_t outr_1 = vqdmulhq_s32(tmp2, factq);
68 int32x4_t outi_0 = vqdmulhq_s32(tmp1, factq)
[all...] |
| entropy_coding_neon.c | 44 int32x4_t shift32x4 = vdupq_n_s32(shift);
46 int32x4_t sum_32x4 = vdupq_n_s32(0);
60 int32x4_t matrix0_32x4 =
62 int32x4_t matrix1_32x4 =
64 int32x4_t multi_32x4 = vqdmulhq_s32(matrix0_32x4, matrix1_32x4);
102 int32x4_t matrix1_32x4 = vdupq_n_s32(matrix1[matrix1_index] << shift);
103 int32x4_t matrix0_32x4 =
105 int32x4_t multi_32x4 = vqdmulhq_s32(matrix0_32x4, matrix1_32x4);
146 int32x4_t matrix1_32x4 =
148 int32x4_t matrix0_32x4
[all...] |
| /external/skia/src/opts/ |
| SkBitmapProcState_matrixProcs_neon.cpp | 22 static inline int16x8_t sbpsm_clamp_tile8(int32x4_t low, int32x4_t high, unsigned max) {
36 static inline int32x4_t sbpsm_clamp_tile4(int32x4_t f, unsigned max) {
37 int32x4_t res;
50 static inline int32x4_t sbpsm_clamp_tile4_low_bits(int32x4_t fx) {
51 int32x4_t ret;
64 static inline int16x8_t sbpsm_repeat_tile8(int32x4_t low, int32x4_t high, unsigned max)
[all...] |
| /prebuilts/gcc/darwin-x86/arm/arm-eabi-4.8/lib/gcc/arm-eabi/4.8/include/ |
| arm_neon.h | 54 typedef __builtin_neon_si int32x4_t __attribute__ ((__vector_size__ (16))); typedef
95 int32x4_t val[2];
205 int32x4_t val[3];
315 int32x4_t val[4];
465 __extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
466 vaddq_s32 (int32x4_t __a, int32x4_t __b)
468 return (int32x4_t)__builtin_neon_vaddv4si (__a, __b, 1);
498 return (uint32x4_t)__builtin_neon_vaddv4si ((int32x4_t) __a, (int32x4_t) __b, 0)
[all...] |
| /prebuilts/gcc/linux-x86/arm/arm-eabi-4.8/lib/gcc/arm-eabi/4.8/include/ |
| arm_neon.h | 54 typedef __builtin_neon_si int32x4_t __attribute__ ((__vector_size__ (16))); typedef
95 int32x4_t val[2];
205 int32x4_t val[3];
315 int32x4_t val[4];
465 __extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
466 vaddq_s32 (int32x4_t __a, int32x4_t __b)
468 return (int32x4_t)__builtin_neon_vaddv4si (__a, __b, 1);
498 return (uint32x4_t)__builtin_neon_vaddv4si ((int32x4_t) __a, (int32x4_t) __b, 0)
[all...] |
| /external/libhevc/common/arm/ |
| ihevc_weighted_pred_neon_intr.c | 118 int32x4_t i4_tmp1_t;
119 int32x4_t i4_tmp2_t;
120 int32x4_t sto_res_tmp1;
124 int32x4_t tmp_lvl_shift_t;
126 int32x4_t tmp_shift_t;
251 int32x4_t i4_tmp1_t;
252 int32x4_t i4_tmp2_t;
253 int32x4_t sto_res_tmp1;
257 int32x4_t tmp_lvl_shift_t_u, tmp_lvl_shift_t_v;
260 int32x4_t tmp_shift_t
[all...] |
| /external/webrtc/webrtc/common_audio/signal_processing/ |
| min_max_operations_neon.c | 87 int32x4_t in32x4_0 = vld1q_s32(p_start);
89 int32x4_t in32x4_1 = vld1q_s32(p_start);
172 int32x4_t max32x4_0 = vdupq_n_s32(WEBRTC_SPL_WORD32_MIN);
173 int32x4_t max32x4_1 = vdupq_n_s32(WEBRTC_SPL_WORD32_MIN);
177 int32x4_t in32x4_0 = vld1q_s32(p_start);
179 int32x4_t in32x4_1 = vld1q_s32(p_start);
185 int32x4_t max32x4 = vmaxq_s32(max32x4_0, max32x4_1);
252 int32x4_t min32x4_0 = vdupq_n_s32(WEBRTC_SPL_WORD32_MAX);
253 int32x4_t min32x4_1 = vdupq_n_s32(WEBRTC_SPL_WORD32_MAX);
257 int32x4_t in32x4_0 = vld1q_s32(p_start)
[all...] |
| cross_correlation_neon.c | 30 int32x4_t tmp0 = vmull_s16(vget_low_s16(seq1_16x8),
32 int32x4_t tmp1 = vmull_high_s16(seq1_16x8, seq2_16x8);
34 int32x4_t tmp0 = vmull_s16(vget_low_s16(seq1_16x8),
36 int32x4_t tmp1 = vmull_s16(vget_high_s16(seq1_16x8),
|
| /external/opencv3/3rdparty/libwebp/dsp/ |
| upsampling_neon.c | 110 int32x4_t vrl = vqdmlal_lane_s16(vshll_n_s16(vget_low_s16(vv), 1), \
112 int32x4_t vrh = vqdmlal_lane_s16(vshll_n_s16(vget_high_s16(vv), 1), \
117 int32x4_t vl = vmovl_s16(vget_low_s16(vv)); \
118 int32x4_t vh = vmovl_s16(vget_high_s16(vv)); \
119 int32x4_t ugl = vmlal_lane_s16(vl, vget_low_s16(uu), cf16, 1); \
120 int32x4_t ugh = vmlal_lane_s16(vh, vget_high_s16(uu), cf16, 1); \
121 int32x4_t gcl = vqdmlal_lane_s16(ugl, vget_low_s16(vv), cf16, 2); \
122 int32x4_t gch = vqdmlal_lane_s16(ugh, vget_high_s16(vv), cf16, 2); \
126 int32x4_t ubl = vqdmlal_lane_s16(vshll_n_s16(vget_low_s16(uu), 1), \
128 int32x4_t ubh = vqdmlal_lane_s16(vshll_n_s16(vget_high_s16(uu), 1),
[all...] |
| /prebuilts/gcc/darwin-x86/arm/arm-linux-androideabi-4.9/lib/gcc/arm-linux-androideabi/4.9/include/ |
| arm_neon.h | 57 typedef __builtin_neon_si int32x4_t __attribute__ ((__vector_size__ (16))); typedef
105 int32x4_t val[2];
231 int32x4_t val[3];
357 int32x4_t val[4];
523 __extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
524 vaddq_s32 (int32x4_t __a, int32x4_t __b)
526 return (int32x4_t)__builtin_neon_vaddv4si (__a, __b, 1);
556 return (uint32x4_t)__builtin_neon_vaddv4si ((int32x4_t) __a, (int32x4_t) __b, 0)
[all...] |
| /prebuilts/gcc/linux-x86/arm/arm-linux-androideabi-4.9/lib/gcc/arm-linux-androideabi/4.9/include/ |
| arm_neon.h | 57 typedef __builtin_neon_si int32x4_t __attribute__ ((__vector_size__ (16))); typedef
105 int32x4_t val[2];
231 int32x4_t val[3];
357 int32x4_t val[4];
523 __extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
524 vaddq_s32 (int32x4_t __a, int32x4_t __b)
526 return (int32x4_t)__builtin_neon_vaddv4si (__a, __b, 1);
556 return (uint32x4_t)__builtin_neon_vaddv4si ((int32x4_t) __a, (int32x4_t) __b, 0)
[all...] |
| /external/clang/test/Sema/ |
| arm-neon-types.c | 24 int32x4_t test4(int32x4_t a, vSInt32 b) {
|
| /external/libvpx/libvpx/vp9/encoder/arm/neon/ |
| vp9_error_neon.c | 31 const int32x4_t err0 = vmull_s16(diff_lo, diff_lo);
32 const int32x4_t err1 = vmlal_s16(err0, diff_hi, diff_hi);
|
| /prebuilts/misc/darwin-x86/analyzer/lib/clang/3.3/include/ |
| arm_neon.h | 42 typedef __attribute__((neon_vector_type(4))) int32_t int32x4_t; typedef
83 int32x4_t val[2];
179 int32x4_t val[3];
275 int32x4_t val[4];
354 __ai int32x4_t vmovl_s16(int16x4_t __a) {
355 return (int32x4_t)__builtin_neon_vmovl_v((int8x8_t)__a, 34); }
367 __ai int32x4_t vmull_s16(int16x4_t __a, int16x4_t __b) {
368 return (int32x4_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)__b, 34); }
398 __ai int32x4_t vabdq_s32(int32x4_t __a, int32x4_t __b)
[all...] |
| /prebuilts/misc/linux-x86/analyzer/lib/clang/3.3/include/ |
| arm_neon.h | 42 typedef __attribute__((neon_vector_type(4))) int32_t int32x4_t; typedef
83 int32x4_t val[2];
179 int32x4_t val[3];
275 int32x4_t val[4];
354 __ai int32x4_t vmovl_s16(int16x4_t __a) {
355 return (int32x4_t)__builtin_neon_vmovl_v((int8x8_t)__a, 34); }
367 __ai int32x4_t vmull_s16(int16x4_t __a, int16x4_t __b) {
368 return (int32x4_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)__b, 34); }
398 __ai int32x4_t vabdq_s32(int32x4_t __a, int32x4_t __b)
[all...] |
| /prebuilts/gcc/darwin-x86/x86/x86_64-linux-android-4.9/lib/gcc/x86_64-linux-android/4.9/include/ |
| arm_neon.h | 144 typedef __m128i int32x4_t; typedef
177 int32x4_t val[2];
245 int32x4_t val[4];
309 int32x4_t val[3];
427 int32x4_t vaddq_s32(int32x4_t a, int32x4_t b); // VADD.I32 q0,q0,q0
436 int32x4_t vaddl_s16(int16x4_t a, int16x4_t b); // VADDL.S16 q0,d0,d0
443 int32x4_t vaddw_s16(int32x4_t a, int16x4_t b); // VADDW.S16 q0,q0,d
[all...] |
