| /third_party/libjpeg-turbo/simd/powerpc/ |
| D | jquanti-altivec.c | 119 __vector short row0, row1, row2, row3, row4, row5, row6, row7, in jsimd_quantize_altivec() local
136 row0 = vec_ld(0, workspace); in jsimd_quantize_altivec()
146 row0s = vec_sra(row0, pw_word_bit_m1); in jsimd_quantize_altivec()
154 row0 = vec_xor(row0, row0s); in jsimd_quantize_altivec()
162 row0 = vec_sub(row0, row0s); in jsimd_quantize_altivec()
180 row0 = vec_add(row0, corr0); in jsimd_quantize_altivec()
198 MULTIPLY(row0, recip0, row0); in jsimd_quantize_altivec()
216 MULTIPLY(row0, scale0, row0); in jsimd_quantize_altivec()
225 row0 = vec_xor(row0, row0s); in jsimd_quantize_altivec()
233 row0 = vec_sub(row0, row0s); in jsimd_quantize_altivec()
[all …]
|
| D | jfdctfst-altivec.c | 92 __vector short row0, row1, row2, row3, row4, row5, row6, row7, in jsimd_fdct_ifast_altivec() local
109 row0 = vec_ld(0, data); in jsimd_fdct_ifast_altivec()
135 tmp0 = vec_add(row0, row7); in jsimd_fdct_ifast_altivec()
136 tmp7 = vec_sub(row0, row7); in jsimd_fdct_ifast_altivec()
|
| D | jfdctint-altivec.c | 184 __vector short row0, row1, row2, row3, row4, row5, row6, row7, in jsimd_fdct_islow_altivec() local
213 row0 = vec_ld(0, data); in jsimd_fdct_islow_altivec()
239 tmp0 = vec_add(row0, row7); in jsimd_fdct_islow_altivec()
240 tmp7 = vec_sub(row0, row7); in jsimd_fdct_islow_altivec()
|
| /third_party/flutter/skia/third_party/externals/libjpeg-turbo/simd/powerpc/ |
| D | jquanti-altivec.c | 119 __vector short row0, row1, row2, row3, row4, row5, row6, row7, in jsimd_quantize_altivec() local
136 row0 = vec_ld(0, workspace); in jsimd_quantize_altivec()
146 row0s = vec_sra(row0, pw_word_bit_m1); in jsimd_quantize_altivec()
154 row0 = vec_xor(row0, row0s); in jsimd_quantize_altivec()
162 row0 = vec_sub(row0, row0s); in jsimd_quantize_altivec()
180 row0 = vec_add(row0, corr0); in jsimd_quantize_altivec()
198 MULTIPLY(row0, recip0, row0); in jsimd_quantize_altivec()
216 MULTIPLY(row0, scale0, row0); in jsimd_quantize_altivec()
225 row0 = vec_xor(row0, row0s); in jsimd_quantize_altivec()
233 row0 = vec_sub(row0, row0s); in jsimd_quantize_altivec()
[all …]
|
| D | jfdctfst-altivec.c | 92 __vector short row0, row1, row2, row3, row4, row5, row6, row7, in jsimd_fdct_ifast_altivec() local
109 row0 = vec_ld(0, data); in jsimd_fdct_ifast_altivec()
135 tmp0 = vec_add(row0, row7); in jsimd_fdct_ifast_altivec()
136 tmp7 = vec_sub(row0, row7); in jsimd_fdct_ifast_altivec()
|
| D | jfdctint-altivec.c | 184 __vector short row0, row1, row2, row3, row4, row5, row6, row7, in jsimd_fdct_islow_altivec() local
213 row0 = vec_ld(0, data); in jsimd_fdct_islow_altivec()
239 tmp0 = vec_add(row0, row7); in jsimd_fdct_islow_altivec()
240 tmp7 = vec_sub(row0, row7); in jsimd_fdct_islow_altivec()
|
| /third_party/flutter/skia/src/opts/ |
| D | SkBitmapProcState_opts.h | 88 __m128i row0 = interpolate_in_x(A0,A1, B0,B1, interlaced_x_weights), in interpolate_in_x_and_y() local
93 return _mm_srli_epi16(_mm_add_epi16(_mm_mullo_epi16(row0, wy0), in interpolate_in_x_and_y()
118 auto row0 = (const uint32_t*)((const uint8_t*)s.fPixmap.addr() + y0 * s.fPixmap.rowBytes()), in S32_alpha_D32_filter_DX() local
141 __m128i AB = interpolate_in_x_and_y(row0[x0[0]], row0[x1[0]], in S32_alpha_D32_filter_DX()
143 row0[x0[1]], row0[x1[1]], in S32_alpha_D32_filter_DX()
148 __m128i CD = interpolate_in_x_and_y(row0[x0[2]], row0[x1[2]], in S32_alpha_D32_filter_DX()
150 row0[x0[3]], row0[x1[3]], in S32_alpha_D32_filter_DX()
173 __m128i A = interpolate_in_x_and_y(row0[x0], row0[x1], in S32_alpha_D32_filter_DX()
199 auto row0 = (const uint32_t*)( (const char*)s.fPixmap.addr() + y0 * s.fPixmap.rowBytes() ), in S32_alpha_D32_filter_DX() local
212 const __m128i a00 = _mm_cvtsi32_si128(row0[x0]), in S32_alpha_D32_filter_DX()
[all …]
|
| /third_party/skia/third_party/externals/libjpeg-turbo/simd/arm/ |
| D | jquanti-neon.c | 57 int16x8_t row0 = in jsimd_convsamp_neon() local
74 vst1q_s16(workspace + 0 * DCTSIZE, row0); in jsimd_convsamp_neon()
105 int16x8_t row0 = vld1q_s16(workspace + (i + 0) * DCTSIZE); in jsimd_quantize_neon() local
124 int16x8_t sign_row0 = vshrq_n_s16(row0, 15); in jsimd_quantize_neon()
129 uint16x8_t abs_row0 = vreinterpretq_u16_s16(vabsq_s16(row0)); in jsimd_quantize_neon()
157 row0 = vcombine_s16(vshrn_n_s32(row0_l, 16), vshrn_n_s32(row0_h, 16)); in jsimd_quantize_neon()
165 row0 = vreinterpretq_s16_u16(vshlq_u16(vreinterpretq_u16_s16(row0), in jsimd_quantize_neon()
175 row0 = veorq_s16(row0, sign_row0); in jsimd_quantize_neon()
176 row0 = vsubq_s16(row0, sign_row0); in jsimd_quantize_neon()
185 vst1q_s16(out_ptr + (i + 0) * DCTSIZE, row0); in jsimd_quantize_neon()
|
| D | jidctred-neon.c | 82 int16x8_t row0 = vld1q_s16(coef_block + 0 * DCTSIZE); in jsimd_idct_2x2_neon() local
96 row0 = vmulq_s16(row0, quant_row0); in jsimd_idct_2x2_neon()
110 int32x4_t tmp10_l = vshll_n_s16(vget_low_s16(row0), CONST_BITS + 2); in jsimd_idct_2x2_neon()
111 int32x4_t tmp10_h = vshll_n_s16(vget_high_s16(row0), CONST_BITS + 2); in jsimd_idct_2x2_neon()
124 row0 = vcombine_s16(vrshrn_n_s32(vaddq_s32(tmp10_l, tmp0_l), CONST_BITS), in jsimd_idct_2x2_neon()
130 int16x8x2_t cols_0246_1357 = vtrnq_s16(row0, row1); in jsimd_idct_2x2_neon()
202 int16x8_t row0 = vld1q_s16(coef_block + 0 * DCTSIZE); in jsimd_idct_4x4_neon() local
213 row0 = vmulq_s16(row0, quant_row0); in jsimd_idct_4x4_neon()
240 int16x8_t dcval = vshlq_n_s16(row0, PASS1_BITS); in jsimd_idct_4x4_neon()
241 row0 = dcval; in jsimd_idct_4x4_neon()
[all …]
|
| D | jidctfst-neon.c | 69 int16x8_t row0 = vld1q_s16(coef_block + 0 * DCTSIZE); in jsimd_idct_ifast_neon() local
81 row0 = vmulq_s16(row0, quant_row0); in jsimd_idct_ifast_neon()
101 int16x8_t dcval = row0; in jsimd_idct_ifast_neon()
113 int16x4_t dcval = vget_low_s16(row0); in jsimd_idct_ifast_neon()
127 int16x4_t tmp0 = vget_high_s16(row0); in jsimd_idct_ifast_neon()
176 row0 = vcombine_s16(dcval, vadd_s16(tmp0, tmp7)); in jsimd_idct_ifast_neon()
188 int16x4_t dcval = vget_high_s16(row0); in jsimd_idct_ifast_neon()
202 int16x4_t tmp0 = vget_low_s16(row0); in jsimd_idct_ifast_neon()
251 row0 = vcombine_s16(vadd_s16(tmp0, tmp7), dcval); in jsimd_idct_ifast_neon()
272 int16x8_t tmp0 = row0; in jsimd_idct_ifast_neon()
[all …]
|
| D | jidctint-neon.c | 97 static INLINE void jsimd_idct_islow_pass1_regular(int16x4_t row0,
116 static INLINE void jsimd_idct_islow_pass1_sparse(int16x4_t row0,
201 int16x4_t row0 = vld1_s16(coef_block + 0 * DCTSIZE); in jsimd_idct_islow_neon() local
233 int16x4_t dcval = vshl_n_s16(vmul_s16(row0, quant_row0), PASS1_BITS); in jsimd_idct_islow_neon()
239 jsimd_idct_islow_pass1_sparse(row0, row1, row2, row3, quant_row0, in jsimd_idct_islow_neon()
244 jsimd_idct_islow_pass1_regular(row0, row1, row2, row3, row4, row5, in jsimd_idct_islow_neon()
254 row0 = vld1_s16(coef_block + 0 * DCTSIZE + 4); in jsimd_idct_islow_neon()
287 bitmap = vorr_s16(bitmap, row0); in jsimd_idct_islow_neon()
291 int16x4_t dcval = vshl_n_s16(vmul_s16(row0, quant_row0), PASS1_BITS); in jsimd_idct_islow_neon()
299 jsimd_idct_islow_pass1_sparse(row0, row1, row2, row3, quant_row0, in jsimd_idct_islow_neon()
[all …]
|
| D | jfdctfst-neon.c | 152 int16x8_t row0 = vreinterpretq_s16_s32(rows_04.val[0]); in jsimd_fdct_ifast_neon() local
163 tmp0 = vaddq_s16(row0, row7); in jsimd_fdct_ifast_neon()
164 tmp7 = vsubq_s16(row0, row7); in jsimd_fdct_ifast_neon()
178 row0 = vaddq_s16(tmp10, tmp11); /* phase 3 */ in jsimd_fdct_ifast_neon()
206 vst1q_s16(data + 0 * DCTSIZE, row0); in jsimd_fdct_ifast_neon()
|
| /third_party/skia/src/opts/ |
| D | SkBitmapProcState_opts.h | 53 const uint32_t* row0 = s.fPixmap.addr32(0,y0); in S32_alpha_D32_filter_DX() local
82 skvx::Vec<8,uint32_t> tl = gather(row0, x0), tr = gather(row0, x1), in S32_alpha_D32_filter_DX()
240 auto row0 = (const uint32_t*)((const uint8_t*)s.fPixmap.addr() + y0 * s.fPixmap.rowBytes()), in S32_alpha_D32_filter_DX() local
268 __m128i AB = interpolate_in_x_and_y(row0[x0[A]], row0[x1[A]], in S32_alpha_D32_filter_DX()
270 row0[x0[B]], row0[x1[B]], in S32_alpha_D32_filter_DX()
275 __m128i CD = interpolate_in_x_and_y(row0[x0[C]], row0[x1[C]], in S32_alpha_D32_filter_DX()
277 row0[x0[D]], row0[x1[D]], in S32_alpha_D32_filter_DX()
299 __m128i A = interpolate_in_x_and_y(row0[x0], row0[x1], in S32_alpha_D32_filter_DX()
323 auto row0 = (const uint32_t*)( (const char*)s.fPixmap.addr() + y0 * s.fPixmap.rowBytes() ), in S32_alpha_D32_filter_DX() local
338 const __m128i tl = _mm_cvtsi32_si128(row0[x0]), tr = _mm_cvtsi32_si128(row0[x1]), in S32_alpha_D32_filter_DX()
[all …]
|
| /third_party/skia/third_party/externals/libjpeg-turbo/simd/arm/aarch32/ |
| D | jchuff-neon.c | 52 int16x8_t row0 = vdupq_n_s16(block[0] - last_dc_val); in jsimd_huff_encode_one_block_neon() local
53 row0 = vld1q_lane_s16(block + 1, row0, 1); in jsimd_huff_encode_one_block_neon()
54 row0 = vld1q_lane_s16(block + 8, row0, 2); in jsimd_huff_encode_one_block_neon()
55 row0 = vld1q_lane_s16(block + 16, row0, 3); in jsimd_huff_encode_one_block_neon()
56 row0 = vld1q_lane_s16(block + 9, row0, 4); in jsimd_huff_encode_one_block_neon()
57 row0 = vld1q_lane_s16(block + 2, row0, 5); in jsimd_huff_encode_one_block_neon()
58 row0 = vld1q_lane_s16(block + 3, row0, 6); in jsimd_huff_encode_one_block_neon()
59 row0 = vld1q_lane_s16(block + 10, row0, 7); in jsimd_huff_encode_one_block_neon()
88 int16x8_t abs_row0 = vabsq_s16(row0); in jsimd_huff_encode_one_block_neon()
114 vshlq_u16(vreinterpretq_u16_s16(vshrq_n_s16(row0, 15)), in jsimd_huff_encode_one_block_neon()
|
| /third_party/skia/third_party/externals/swiftshader/src/Pipeline/ |
| D | ShaderCore.cpp | 472 void transpose4x4(Short4 &row0, Short4 &row1, Short4 &row2, Short4 &row3) in transpose4x4() argument
474 Int2 tmp0 = UnpackHigh(row0, row1); in transpose4x4()
476 Int2 tmp2 = UnpackLow(row0, row1); in transpose4x4()
479 row0 = UnpackLow(tmp2, tmp3); in transpose4x4()
485 void transpose4x3(Short4 &row0, Short4 &row1, Short4 &row2, Short4 &row3) in transpose4x3() argument
487 Int2 tmp0 = UnpackHigh(row0, row1); in transpose4x3()
489 Int2 tmp2 = UnpackLow(row0, row1); in transpose4x3()
492 row0 = UnpackLow(tmp2, tmp3); in transpose4x3()
497 void transpose4x4(Float4 &row0, Float4 &row1, Float4 &row2, Float4 &row3) in transpose4x4() argument
499 Float4 tmp0 = UnpackLow(row0, row1); in transpose4x4()
[all …]
|
| /third_party/skia/docs/examples/ |
| D | Bitmap_getAddr.cpp | 7 char* row0 = (char* ) source.getAddr(0, 0); in draw() local
10 (size_t) (row1 - row0) == source.rowBytes() ? '=' : '!'); in draw()
|
| D | Bitmap_getAddr32.cpp | 7 uint32_t* row0 = source.getAddr32(0, 0); in draw() local
9 size_t interval = (row1 - row0) * source.bytesPerPixel(); in draw()
|
| D | Bitmap_getAddr16.cpp | 12 uint16_t* row0 = bitmap16.getAddr16(0, 0); in draw() local
14 size_t interval = (row1 - row0) * bitmap16.bytesPerPixel(); in draw()
|
| /third_party/flutter/skia/docs/examples/ |
| D | Bitmap_getAddr.cpp | 7 char* row0 = (char* ) source.getAddr(0, 0); in draw() local
10 (size_t) (row1 - row0) == source.rowBytes() ? '=' : '!'); in draw()
|
| D | Bitmap_getAddr32.cpp | 7 uint32_t* row0 = source.getAddr32(0, 0); in draw() local
9 size_t interval = (row1 - row0) * source.bytesPerPixel(); in draw()
|
| D | Bitmap_getAddr16.cpp | 12 uint16_t* row0 = bitmap16.getAddr16(0, 0); in draw() local
14 size_t interval = (row1 - row0) * bitmap16.bytesPerPixel(); in draw()
|
| /third_party/ffmpeg/libavcodec/mips/ |
| D | vp8_lpf_msa.c | 380 v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8; in ff_vp8_h_loop_filter16_msa() local
388 LD_UB8(temp_src, pitch, row0, row1, row2, row3, row4, row5, row6, row7); in ff_vp8_h_loop_filter16_msa()
392 TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7, in ff_vp8_h_loop_filter16_msa()
445 v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8; in ff_vp8_h_loop_filter8uv_msa() local
453 LD_UB8(src_u - 4, pitch, row0, row1, row2, row3, row4, row5, row6, row7); in ff_vp8_h_loop_filter8uv_msa()
456 TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7, in ff_vp8_h_loop_filter8uv_msa()
525 v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8; in ff_vp8_h_loop_filter_simple_msa() local
531 LD_UB8(temp_src, pitch, row0, row1, row2, row3, row4, row5, row6, row7); in ff_vp8_h_loop_filter_simple_msa()
535 TRANSPOSE16x4_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7, in ff_vp8_h_loop_filter_simple_msa()
595 v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8; in ff_vp8_h_loop_filter8uv_inner_msa() local
[all …]
|
| /third_party/skia/third_party/externals/swiftshader/src/Shader/ |
| D | ShaderCore.hpp | 87 void transpose4x4(Short4 &row0, Short4 &row1, Short4 &row2, Short4 &row3);
88 void transpose4x3(Short4 &row0, Short4 &row1, Short4 &row2, Short4 &row3);
89 void transpose4x4(Float4 &row0, Float4 &row1, Float4 &row2, Float4 &row3);
90 void transpose4x3(Float4 &row0, Float4 &row1, Float4 &row2, Float4 &row3);
91 void transpose4x2(Float4 &row0, Float4 &row1, Float4 &row2, Float4 &row3);
92 void transpose4x1(Float4 &row0, Float4 &row1, Float4 &row2, Float4 &row3);
93 void transpose2x4(Float4 &row0, Float4 &row1, Float4 &row2, Float4 &row3);
94 void transpose4xN(Float4 &row0, Float4 &row1, Float4 &row2, Float4 &row3, int N);
|
| /third_party/mesa3d/src/gallium/drivers/swr/rasterizer/core/ |
| D | format_utils.h | 34 void vTranspose(simd4scalar& row0, simd4scalar& row1, simd4scalar& row2, simd4scalar& row3) in vTranspose() argument
36 simd4scalari row0i = SIMD128::castps_si(row0); in vTranspose()
57 row0 = SIMD128::castsi_ps(row0i); in vTranspose()
64 void vTranspose(simd4scalari& row0, simd4scalari& row1, simd4scalari& row2, simd4scalari& row3) in vTranspose() argument
70 row3 = row0; in vTranspose()
71 row0 = SIMD128::unpacklo_epi32(row0, row1); in vTranspose()
74 row1 = row0; in vTranspose()
75 row0 = SIMD128::unpacklo_epi64(row0, row2); in vTranspose()
|
| /third_party/skia/third_party/externals/libjpeg-turbo/simd/arm/aarch64/ |
| D | jchuff-neon.c | 121 int16x8_t row0 = in jsimd_huff_encode_one_block_neon() local
139 row0 = vsetq_lane_s16(block[0] - last_dc_val, row0, 0); in jsimd_huff_encode_one_block_neon()
161 int16x8_t abs_row0 = vabsq_s16(row0); in jsimd_huff_encode_one_block_neon()
172 vreinterpretq_u16_s16(veorq_s16(abs_row0, vshrq_n_s16(row0, 15))); in jsimd_huff_encode_one_block_neon()
|